OpenAI's hidden model claims 372 math results | The Austin Wilson Show, Oct 7, 2026
2026-10-07
https://www.youtube.com/watch?v=FEPXlTM2Mi0

[0:21] We go live in 10.
[0:35] Every morning, the future moves faster.
[0:40] Every morning, the future moves faster. The models, the money, the power, and
[0:44] The models, the money, the power, and
[0:44] The models, the money, the power, and the people building what comes next.
[1:19] from Austin, [music] Texas to the world.
[1:26] AI news, real context. This is the
[1:30] AI news, real context. This is the Austin Wilson show. [music]
[1:32] Austin Wilson show. [music]
[1:32] Austin Wilson show. [music] I'm Austin Wilson, your human in the
[1:34] I'm Austin Wilson, your human in the
[1:34] I'm Austin Wilson, your human in the loop.
[1:41] Welcome to the Austin Wilson Show, where
[1:44] Welcome to the Austin Wilson Show, where we're breaking the AI headlines that
[1:46] we're breaking the AI headlines that
[1:46] we're breaking the AI headlines that matter to you today. I'm Austin Wilson
[1:49] matter to you today. I'm Austin Wilson
[1:49] matter to you today. I'm Austin Wilson coming to you live from the AI capital
[1:51] coming to you live from the AI capital
[1:51] coming to you live from the AI capital of the world, Austin, Texas. Today,
[1:54] of the world, Austin, Texas. Today,
[1:54] of the world, Austin, Texas. Today, OpenAI's hidden model claims to solve
[1:56] OpenAI's hidden model claims to solve
[1:56] OpenAI's hidden model claims to solve nearly 400 math problems. SpaceX is
[2:00] nearly 400 math problems. SpaceX is
[2:00] nearly 400 math problems. SpaceX is reportedly borrowing $40 billion. Meta,
[2:04] reportedly borrowing $40 billion. Meta,
[2:04] reportedly borrowing $40 billion. Meta, Walmart, Stripe write rules for AI
[2:07] Walmart, Stripe write rules for AI
[2:07] Walmart, Stripe write rules for AI shoppers. Europe's Misrel and a one
[2:10] shoppers. Europe's Misrel and a one
[2:10] shoppers. Europe's Misrel and a one trillion parameter open model. Lambda
[2:13] trillion parameter open model. Lambda
[2:13] trillion parameter open model. Lambda reportedly raising $4 billion before its
[2:16] reportedly raising $4 billion before its
[2:16] reportedly raising $4 billion before its IPO. and Google upgrades its nano banana
[2:21] IPO. and Google upgrades its nano banana
[2:21] IPO. and Google upgrades its nano banana image AI. Let's get into it. First up,
[2:24] image AI. Let's get into it. First up,
[2:24] image AI. Let's get into it. First up, this is amazing. Open AAI says an
[2:27] this is amazing. Open AAI says an
[2:27] this is amazing. Open AAI says an unreleased model produced 372 new math
[2:30] unreleased model produced 372 new math
[2:30] unreleased model produced 372 new math results, including a claimed solution to
[2:33] results, including a claimed solution to
[2:33] results, including a claimed solution to the four-dimensional kia conjecture,
[2:35] the four-dimensional kia conjecture,
[2:36] the four-dimensional kia conjecture, nearly all from a single prompt. OpenAI
[2:39] nearly all from a single prompt. OpenAI
[2:39] nearly all from a single prompt. OpenAI posted 722 manuscripts and 372 families
[2:43] posted 722 manuscripts and 372 families
[2:43] posted 722 manuscripts and 372 families to GitHub Tuesday evening and says each
[2:47] to GitHub Tuesday evening and says each
[2:47] to GitHub Tuesday evening and says each one resolves or makes real progress on
[2:49] one resolves or makes real progress on
[2:49] one resolves or makes real progress on an open question in math or theoretical
[2:53] an open question in math or theoretical
[2:53] an open question in math or theoretical computer science. The model was given
[2:55] computer science. The model was given
[2:55] computer science. The model was given about 4,000 problems and the average
[2:57] about 4,000 problems and the average
[2:57] about 4,000 problems and the average result used about 3 hours of chat GBT
[3:01] result used about 3 hours of chat GBT
[3:01] result used about 3 hours of chat GBT pro thinking last month. It's Navier
[3:04] pro thinking last month. It's Navier
[3:04] pro thinking last month. It's Navier Stokes claim took a swarm of about
[3:07] Stokes claim took a swarm of about
[3:07] Stokes claim took a swarm of about 10,000 agents. This time it says one
[3:09] 10,000 agents. This time it says one
[3:09] 10,000 agents. This time it says one agent nearly did all of it.
[3:14] agent nearly did all of it.
[3:14] agent nearly did all of it. If it holds up, AI is now producing
[3:17] If it holds up, AI is now producing
[3:17] If it holds up, AI is now producing research results faster than
[3:19] research results faster than
[3:20] research results faster than mathematicians can check them. Fields
[3:22] mathematicians can check them. Fields
[3:22] mathematicians can check them. Fields medalist Terrence Tao has criticized the
[3:25] medalist Terrence Tao has criticized the
[3:25] medalist Terrence Tao has criticized the pace. OpenAI says nearly every result
[3:28] pace. OpenAI says nearly every result
[3:28] pace. OpenAI says nearly every result came from one prompt to one agent and
[3:31] came from one prompt to one agent and
[3:31] came from one prompt to one agent and says the model is coming soon. Many of
[3:34] says the model is coming soon. Many of
[3:34] says the model is coming soon. Many of the proofs are checked in lean, a
[3:37] the proofs are checked in lean, a
[3:37] the proofs are checked in lean, a proof-checking language which Scientific
[3:39] proof-checking language which Scientific
[3:39] proof-checking language which Scientific American says makes them all but certain
[3:42] American says makes them all but certain
[3:42] American says makes them all but certain to be correct. Uh OpenAI hasn't released
[3:46] to be correct. Uh OpenAI hasn't released
[3:46] to be correct. Uh OpenAI hasn't released the model, the exact prompts or the
[3:48] the model, the exact prompts or the
[3:48] the model, the exact prompts or the compute per problem. MIT's Andrew
[3:51] compute per problem. MIT's Andrew
[3:51] compute per problem. MIT's Andrew Sutherland says to treat claims of
[3:53] Sutherland says to treat claims of
[3:53] Sutherland says to treat claims of one-shotting problems with a single uh
[3:56] one-shotting problems with a single uh
[3:56] one-shotting problems with a single uh agent as unverified until they've
[4:00] agent as unverified until they've
[4:00] agent as unverified until they've actually been tested by humans.
[4:04] actually been tested by humans.
[4:04] actually been tested by humans. Man, uh this is a phenomenal advancement
[4:08] Man, uh this is a phenomenal advancement
[4:08] Man, uh this is a phenomenal advancement and and an amazing release from Open AI,
[4:11] and and an amazing release from Open AI,
[4:11] and and an amazing release from Open AI, maybe to counter all of their um
[4:13] maybe to counter all of their um
[4:14] maybe to counter all of their um security wos that they've been having
[4:15] security wos that they've been having
[4:15] security wos that they've been having recently. uh they released 372
[4:21] recently. uh they released 372
[4:21] recently. uh they released 372 um claims to to solve
[4:25] um claims to to solve
[4:25] um claims to to solve some very very um
[4:28] some very very um
[4:28] some very very um like very u major math problems. Right?
[4:32] like very u major math problems. Right?
[4:32] like very u major math problems. Right? So these a lot of these problems um
[4:35] So these a lot of these problems um
[4:35] So these a lot of these problems um humans have been working on for decades
[4:37] humans have been working on for decades
[4:37] humans have been working on for decades and even up to a century in some case
[4:39] and even up to a century in some case
[4:39] and even up to a century in some case where we have been trying to make
[4:42] where we have been trying to make
[4:42] where we have been trying to make progress on these have not made little
[4:45] progress on these have not made little
[4:45] progress on these have not made little progress um uh you know little solves
[4:49] progress um uh you know little solves
[4:49] progress um uh you know little solves here or there or incremental progress
[4:51] here or there or incremental progress
[4:51] here or there or incremental progress but no major breakthroughs. Uh and so
[4:55] but no major breakthroughs. Uh and so
[4:55] but no major breakthroughs. Uh and so for open AAI to release uh claims to
[4:59] for open AAI to release uh claims to
[4:59] for open AAI to release uh claims to solutions for 372 of these is
[5:02] solutions for 372 of these is
[5:02] solutions for 372 of these is mindblowing. Now uh the direct
[5:06] mindblowing. Now uh the direct
[5:06] mindblowing. Now uh the direct application and what does this matter to
[5:08] application and what does this matter to
[5:08] application and what does this matter to the everyday person? Does this change
[5:10] the everyday person? Does this change
[5:10] the everyday person? Does this change anything immediately or what's the
[5:13] anything immediately or what's the
[5:13] anything immediately or what's the direct application? There really isn't a
[5:17] direct application? There really isn't a
[5:17] direct application? There really isn't a lot of direct application here. It's
[5:19] lot of direct application here. It's
[5:19] lot of direct application here. It's it's not like a math problem that they
[5:21] it's not like a math problem that they
[5:21] it's not like a math problem that they solved is somehow going to make our cars
[5:23] solved is somehow going to make our cars
[5:23] solved is somehow going to make our cars more efficient or um you know make
[5:27] more efficient or um you know make
[5:27] more efficient or um you know make computers that much smarter or anything
[5:29] computers that much smarter or anything
[5:29] computers that much smarter or anything like that. The the key breakthrough here
[5:31] like that. The the key breakthrough here
[5:31] like that. The the key breakthrough here is kind of twofold. Number one, uh the
[5:34] is kind of twofold. Number one, uh the
[5:34] is kind of twofold. Number one, uh the fact that the the AI models are able to
[5:38] fact that the the AI models are able to
[5:38] fact that the the AI models are able to solve these problems is a huge just uh
[5:42] solve these problems is a huge just uh
[5:42] solve these problems is a huge just uh breakthrough from an understanding
[5:44] breakthrough from an understanding
[5:44] breakthrough from an understanding perspective that they are able to uh
[5:48] perspective that they are able to uh
[5:48] perspective that they are able to uh solve at this volume and at this
[5:49] solve at this volume and at this
[5:50] solve at this volume and at this difficulty is really important moving
[5:52] difficulty is really important moving
[5:52] difficulty is really important moving forward and just opens up the uh kind of
[5:55] forward and just opens up the uh kind of
[5:55] forward and just opens up the uh kind of aperture of the the possibility of what
[5:59] aperture of the the possibility of what
[5:59] aperture of the the possibility of what AI can go solve in the future. and and
[6:02] AI can go solve in the future. and and
[6:02] AI can go solve in the future. and and the fact that these are novel solutions.
[6:05] the fact that these are novel solutions.
[6:05] the fact that these are novel solutions. So this is not a human uh you know
[6:08] So this is not a human uh you know
[6:08] So this is not a human uh you know saying write me a poem for my niece's
[6:11] saying write me a poem for my niece's
[6:11] saying write me a poem for my niece's birthday and and the AI comes up with a
[6:14] birthday and and the AI comes up with a
[6:14] birthday and and the AI comes up with a cool rhyming thing. No, this is like uh
[6:17] cool rhyming thing. No, this is like uh
[6:17] cool rhyming thing. No, this is like uh they pointed it at 4,000 different
[6:19] they pointed it at 4,000 different
[6:19] they pointed it at 4,000 different problems and said go figure out which
[6:20] problems and said go figure out which
[6:20] problems and said go figure out which ones you can solve. Right? So that's a
[6:22] ones you can solve. Right? So that's a
[6:22] ones you can solve. Right? So that's a that's a big breakthrough from that
[6:24] that's a big breakthrough from that
[6:24] that's a big breakthrough from that standpoint. The other notable thing here
[6:27] standpoint. The other notable thing here
[6:27] standpoint. The other notable thing here is uh the fact that it took on average
[6:31] is uh the fact that it took on average
[6:31] is uh the fact that it took on average three hours of uh a chat GBT pro which
[6:37] three hours of uh a chat GBT pro which
[6:37] three hours of uh a chat GBT pro which is a consumer grade model um you know 3
[6:41] is a consumer grade model um you know 3
[6:41] is a consumer grade model um you know 3 hours of prolevel thinking. Of course,
[6:43] hours of prolevel thinking. Of course,
[6:43] hours of prolevel thinking. Of course, this is an unreleased model, so this
[6:45] this is an unreleased model, so this
[6:45] this is an unreleased model, so this model is going to be more advanced, but
[6:47] model is going to be more advanced, but
[6:47] model is going to be more advanced, but just the perceived compute power that it
[6:50] just the perceived compute power that it
[6:50] just the perceived compute power that it took to solve each of these um is very
[6:53] took to solve each of these um is very
[6:53] took to solve each of these um is very small, especially, you know, when we
[6:55] small, especially, you know, when we
[6:55] small, especially, you know, when we look back at the Navier Stokes solution
[6:59] look back at the Navier Stokes solution
[6:59] look back at the Navier Stokes solution um you know, just several weeks ago um
[7:02] um you know, just several weeks ago um
[7:02] um you know, just several weeks ago um that took 10,000 agents in 88 hours uh
[7:06] that took 10,000 agents in 88 hours uh
[7:06] that took 10,000 agents in 88 hours uh to solve. And now we're seeing agents
[7:09] to solve. And now we're seeing agents
[7:09] to solve. And now we're seeing agents solve these much faster which with much
[7:12] solve these much faster which with much
[7:12] solve these much faster which with much less compute which uh in and of itself
[7:15] less compute which uh in and of itself
[7:15] less compute which uh in and of itself is a phenomenal breakthrough. And then
[7:16] is a phenomenal breakthrough. And then
[7:16] is a phenomenal breakthrough. And then and then maybe the third piece that goes
[7:19] and then maybe the third piece that goes
[7:19] and then maybe the third piece that goes with this is just seeing that yeah
[7:21] with this is just seeing that yeah
[7:21] with this is just seeing that yeah Navier Stokes was maybe a month ago or a
[7:23] Navier Stokes was maybe a month ago or a
[7:23] Navier Stokes was maybe a month ago or a little over a month ago and um and we've
[7:27] little over a month ago and um and we've
[7:27] little over a month ago and um and we've advanced this much in that short of
[7:29] advanced this much in that short of
[7:29] advanced this much in that short of period of time and so just the rate of
[7:31] period of time and so just the rate of
[7:31] period of time and so just the rate of advancement is another phenomenal thing.
[7:33] advancement is another phenomenal thing.
[7:33] advancement is another phenomenal thing. So, uh, you know, big congrats to Open
[7:37] So, uh, you know, big congrats to Open
[7:37] So, uh, you know, big congrats to Open AAI and this release here and we hope to
[7:39] AAI and this release here and we hope to
[7:39] AAI and this release here and we hope to see more of this. This is really
[7:41] see more of this. This is really
[7:41] see more of this. This is really exciting. Um, and and look forward to
[7:44] exciting. Um, and and look forward to
[7:44] exciting. Um, and and look forward to seeing kind of the solutions that
[7:46] seeing kind of the solutions that
[7:46] seeing kind of the solutions that actually do have a positive impact on
[7:49] actually do have a positive impact on
[7:49] actually do have a positive impact on the day-to-day.
[7:51] the day-to-day.
[7:51] the day-to-day. SpaceX is reportedly lining up $40
[7:53] SpaceX is reportedly lining up $40
[7:53] SpaceX is reportedly lining up $40 billion of debt led by Apollo. Uh, this
[7:57] billion of debt led by Apollo. Uh, this
[7:57] billion of debt led by Apollo. Uh, this is an order to buy Nvidia chips for its
[7:59] is an order to buy Nvidia chips for its
[8:00] is an order to buy Nvidia chips for its AI data centers. The Financial Times
[8:02] AI data centers. The Financial Times
[8:02] AI data centers. The Financial Times reported Tuesday that the roughly $40
[8:05] reported Tuesday that the roughly $40
[8:05] reported Tuesday that the roughly $40 billion package would be about $10
[8:07] billion package would be about $10
[8:07] billion package would be about $10 billion of bank loans and $30 billion of
[8:11] billion of bank loans and $30 billion of
[8:11] billion of bank loans and $30 billion of investment grade bonds with Apollo
[8:14] investment grade bonds with Apollo
[8:14] investment grade bonds with Apollo organizing it and bond giant PIMCO among
[8:17] organizing it and bond giant PIMCO among
[8:17] organizing it and bond giant PIMCO among the firms in the talks. It may not close
[8:19] the firms in the talks. It may not close
[8:19] the firms in the talks. It may not close until 2027.
[8:22] until 2027.
[8:22] until 2027. Old Elon Musk has said XAI's Colossus 2
[8:25] Old Elon Musk has said XAI's Colossus 2
[8:25] Old Elon Musk has said XAI's Colossus 2 data center could double its Nvidia
[8:27] data center could double its Nvidia
[8:27] data center could double its Nvidia chips by December. Uh that's amazing. AI
[8:32] chips by December. Uh that's amazing. AI
[8:32] chips by December. Uh that's amazing. AI is now being built on borrowed money or
[8:35] is now being built on borrowed money or
[8:35] is now being built on borrowed money or continues to be built on borrowed money
[8:37] continues to be built on borrowed money
[8:37] continues to be built on borrowed money like railroads and power plants uh and a
[8:40] like railroads and power plants uh and a
[8:40] like railroads and power plants uh and a lot of the internet build out and we've
[8:43] lot of the internet build out and we've
[8:43] lot of the internet build out and we've see this across uh um technological
[8:46] see this across uh um technological
[8:46] see this across uh um technological revolutions. Uh Morgan Stanley estimates
[8:49] revolutions. Uh Morgan Stanley estimates
[8:49] revolutions. Uh Morgan Stanley estimates AI infrastructure needs about $1.5
[8:52] AI infrastructure needs about $1.5
[8:52] AI infrastructure needs about $1.5 trillion of outside financing by 2028.
[8:55] trillion of outside financing by 2028.
[8:56] trillion of outside financing by 2028. Reuters reports SpaceX plans to use
[8:58] Reuters reports SpaceX plans to use
[8:58] Reuters reports SpaceX plans to use Nvidia chips exclusively in its future
[9:01] Nvidia chips exclusively in its future
[9:02] Nvidia chips exclusively in its future data centers. That's interesting. Per
[9:04] data centers. That's interesting. Per
[9:04] data centers. That's interesting. Per Reuters, SpaceX shares fell 1%. Nvidia
[9:07] Reuters, SpaceX shares fell 1%. Nvidia
[9:07] Reuters, SpaceX shares fell 1%. Nvidia rose half a percent in extended trading.
[9:10] rose half a percent in extended trading.
[9:10] rose half a percent in extended trading. On the report, debt gets paid back
[9:12] On the report, debt gets paid back
[9:12] On the report, debt gets paid back whether or not the AI pays off. And this
[9:15] whether or not the AI pays off. And this
[9:15] whether or not the AI pays off. And this deal may not even close until 2027.
[9:18] deal may not even close until 2027.
[9:18] deal may not even close until 2027. We love to see it. We love to see
[9:20] We love to see it. We love to see
[9:20] We love to see it. We love to see putting the chips on the table, maybe
[9:22] putting the chips on the table, maybe
[9:22] putting the chips on the table, maybe borrowing some chips to put on the
[9:23] borrowing some chips to put on the
[9:23] borrowing some chips to put on the table. This is uh an amazing from um the
[9:29] table. This is uh an amazing from um the
[9:29] table. This is uh an amazing from um the the Elon companies from SpaceX AI
[9:31] the Elon companies from SpaceX AI
[9:31] the Elon companies from SpaceX AI buildout.
[9:33] buildout.
[9:33] buildout. To give some context here, um Elon Musk
[9:37] To give some context here, um Elon Musk
[9:37] To give some context here, um Elon Musk and the the SpaceX AI team uh went and
[9:42] and the the SpaceX AI team uh went and
[9:42] and the the SpaceX AI team uh went and build out built out their initial data
[9:44] build out built out their initial data
[9:44] build out built out their initial data center Colossus. I believe it's up to a
[9:47] center Colossus. I believe it's up to a
[9:48] center Colossus. I believe it's up to a 200,000 GPU cluster which is massive
[9:52] 200,000 GPU cluster which is massive
[9:52] 200,000 GPU cluster which is massive massive massive uh amount of compute. Um
[9:56] massive massive uh amount of compute. Um
[9:56] massive massive uh amount of compute. Um they they of course did not uh launch
[9:59] they they of course did not uh launch
[9:59] they they of course did not uh launch the data center at that scale. Um it it
[10:02] the data center at that scale. Um it it
[10:02] the data center at that scale. Um it it took a little bit of time to to build up
[10:03] took a little bit of time to to build up
[10:03] took a little bit of time to to build up to that but they were able to to start
[10:06] to that but they were able to to start
[10:06] to that but they were able to to start from nothing and have a data data center
[10:09] from nothing and have a data data center
[10:09] from nothing and have a data data center in a very very short period of time. I
[10:11] in a very very short period of time. I
[10:11] in a very very short period of time. I believe it was 162 days or or somewhere
[10:14] believe it was 162 days or or somewhere
[10:14] believe it was 162 days or or somewhere around that. Don't quote me on that, but
[10:15] around that. Don't quote me on that, but
[10:15] around that. Don't quote me on that, but it was a very short period of time to go
[10:17] it was a very short period of time to go
[10:17] it was a very short period of time to go from uh breaking ground to having a
[10:21] from uh breaking ground to having a
[10:21] from uh breaking ground to having a running data center. And Elon's just
[10:23] running data center. And Elon's just
[10:23] running data center. And Elon's just running at this uh full steam ahead. Um
[10:27] running at this uh full steam ahead. Um
[10:27] running at this uh full steam ahead. Um and seeing this uh debt issuance $40
[10:31] and seeing this uh debt issuance $40
[10:31] and seeing this uh debt issuance $40 billion to go grab more chips and
[10:34] billion to go grab more chips and
[10:34] billion to go grab more chips and continue to build out. Um it it's kind
[10:36] continue to build out. Um it it's kind
[10:36] continue to build out. Um it it's kind of an interesting thing here where we
[10:38] of an interesting thing here where we
[10:38] of an interesting thing here where we see um [snorts] at the the top of the
[10:42] see um [snorts] at the the top of the
[10:42] see um [snorts] at the the top of the stack right now kind of to to explain
[10:44] stack right now kind of to to explain
[10:44] stack right now kind of to to explain this in layman's terms at the top of the
[10:46] this in layman's terms at the top of the
[10:46] this in layman's terms at the top of the stack we have your application. So, um,
[10:49] stack we have your application. So, um,
[10:49] stack we have your application. So, um, you know, maybe think about we're
[10:50] you know, maybe think about we're
[10:50] you know, maybe think about we're starting to see some applications come
[10:52] starting to see some applications come
[10:52] starting to see some applications come out. If you work at a company that uses
[10:55] out. If you work at a company that uses
[10:55] out. If you work at a company that uses Glean as an enterprise search agent or,
[10:58] Glean as an enterprise search agent or,
[10:58] Glean as an enterprise search agent or, um, maybe you're using some of these,
[11:01] um, maybe you're using some of these,
[11:01] um, maybe you're using some of these, uh, personal assistant, uh, agents like
[11:03] uh, personal assistant, uh, agents like
[11:03] uh, personal assistant, uh, agents like Muse or Grockbot. Um, those applications
[11:07] Muse or Grockbot. Um, those applications
[11:07] Muse or Grockbot. Um, those applications are consumerf facing uh, very much at
[11:09] are consumerf facing uh, very much at
[11:10] are consumerf facing uh, very much at the top of the stack. Then we move down
[11:12] the top of the stack. Then we move down
[11:12] the top of the stack. Then we move down to kind of the middle layer. That's
[11:13] to kind of the middle layer. That's
[11:13] to kind of the middle layer. That's where the AI models sit. So we've got uh
[11:17] where the AI models sit. So we've got uh
[11:17] where the AI models sit. So we've got uh open AI and we've got anthropic with
[11:19] open AI and we've got anthropic with
[11:19] open AI and we've got anthropic with cloud. We've got gro XAI, we've got um
[11:23] cloud. We've got gro XAI, we've got um
[11:23] cloud. We've got gro XAI, we've got um open- source models, we've got different
[11:25] open- source models, we've got different
[11:25] open- source models, we've got different models that are being built out and
[11:27] models that are being built out and
[11:27] models that are being built out and those are feeding those applications. So
[11:29] those are feeding those applications. So
[11:29] those are feeding those applications. So of course in order to develop those
[11:31] of course in order to develop those
[11:31] of course in order to develop those models um to train those models to get
[11:34] models um to train those models to get
[11:34] models um to train those models to get smarter and smarter models as well as to
[11:37] smarter and smarter models as well as to
[11:37] smarter and smarter models as well as to serve those models to the the the
[11:40] serve those models to the the the
[11:40] serve those models to the the the front-end user um they need to run on
[11:43] front-end user um they need to run on
[11:44] front-end user um they need to run on data centers and they need to be trained
[11:46] data centers and they need to be trained
[11:46] data centers and they need to be trained and uh and and so it's interesting to
[11:48] and uh and and so it's interesting to
[11:48] and uh and and so it's interesting to see where does the value go in the stack
[11:51] see where does the value go in the stack
[11:51] see where does the value go in the stack because then at the bottom of the layer
[11:53] because then at the bottom of the layer
[11:53] because then at the bottom of the layer we have uh infra uh the infrastructure
[11:55] we have uh infra uh the infrastructure
[11:55] we have uh infra uh the infrastructure right so this is the data centers this
[11:58] right so this is the data centers this
[11:58] right so this is the data centers this is the bet that Elon's making that
[12:00] is the bet that Elon's making that
[12:00] is the bet that Elon's making that Nvidia with their chips are making and
[12:03] Nvidia with their chips are making and
[12:03] Nvidia with their chips are making and that we continue to see a lot of money
[12:05] that we continue to see a lot of money
[12:06] that we continue to see a lot of money move towards the infrastructure and the
[12:08] move towards the infrastructure and the
[12:08] move towards the infrastructure and the compute layer where the actual, you
[12:12] compute layer where the actual, you
[12:12] compute layer where the actual, you know, servers and computers are running
[12:14] know, servers and computers are running
[12:14] know, servers and computers are running to serve the the the rest of the stack.
[12:17] to serve the the the rest of the stack.
[12:17] to serve the the the rest of the stack. And so seeing this bet made by Elon, um
[12:21] And so seeing this bet made by Elon, um
[12:21] And so seeing this bet made by Elon, um it it does two things. Number one, it's
[12:24] it it does two things. Number one, it's
[12:24] it it does two things. Number one, it's like, okay, um Elon's not one for being
[12:28] like, okay, um Elon's not one for being
[12:28] like, okay, um Elon's not one for being reserved when when he takes risks. He
[12:30] reserved when when he takes risks. He
[12:30] reserved when when he takes risks. He takes very big risks. So that that is,
[12:32] takes very big risks. So that that is,
[12:32] takes very big risks. So that that is, you know, something to be considered,
[12:34] you know, something to be considered,
[12:34] you know, something to be considered, but calculated nonetheless. And so we we
[12:37] but calculated nonetheless. And so we we
[12:37] but calculated nonetheless. And so we we should see that um he he must see
[12:40] should see that um he he must see
[12:40] should see that um he he must see something. They their team must see
[12:41] something. They their team must see
[12:41] something. They their team must see something in order for um you know,
[12:45] something in order for um you know,
[12:45] something in order for um you know, Apollo and big financiers to back uh
[12:49] Apollo and big financiers to back uh
[12:49] Apollo and big financiers to back uh this kind of move. they they must
[12:51] this kind of move. they they must
[12:51] this kind of move. they they must believe in the project and and what
[12:54] believe in the project and and what
[12:54] believe in the project and and what they're doing for Colossus 2 in order to
[12:57] they're doing for Colossus 2 in order to
[12:57] they're doing for Colossus 2 in order to back that. And so there is some positive
[12:59] back that. And so there is some positive
[12:59] back that. And so there is some positive signal. There is some um you know
[13:02] signal. There is some um you know
[13:02] signal. There is some um you know accountability here where where we see
[13:04] accountability here where where we see
[13:04] accountability here where where we see hey maybe there is something real here.
[13:07] hey maybe there is something real here.
[13:07] hey maybe there is something real here. May maybe there's something real. I'm
[13:09] May maybe there's something real. I'm
[13:09] May maybe there's something real. I'm pretty sure there's something real here.
[13:10] pretty sure there's something real here.
[13:10] pretty sure there's something real here. $40 billion. That's what I'm talking
[13:12] $40 billion. That's what I'm talking
[13:12] $40 billion. That's what I'm talking about. Um, that being said, uh, we've
[13:15] about. Um, that being said, uh, we've
[13:15] about. Um, that being said, uh, we've seen this in past revolutions,
[13:17] seen this in past revolutions,
[13:17] seen this in past revolutions, technological revolutions with the
[13:19] technological revolutions with the
[13:19] technological revolutions with the railroad build out, most recently with
[13:21] railroad build out, most recently with
[13:21] railroad build out, most recently with with the internet build out. And in
[13:23] with the internet build out. And in
[13:23] with the internet build out. And in specifically in the internet's case, um,
[13:26] specifically in the internet's case, um,
[13:26] specifically in the internet's case, um, everybody,
[13:27] everybody,
[13:28] everybody, you know, really believed and got behind
[13:30] you know, really believed and got behind
[13:30] you know, really believed and got behind uh, networking and laying fiber cable.
[13:33] uh, networking and laying fiber cable.
[13:33] uh, networking and laying fiber cable. Uh, we still use fiber internet today
[13:36] Uh, we still use fiber internet today
[13:36] Uh, we still use fiber internet today and and how do we get the whole world
[13:38] and and how do we get the whole world
[13:38] and and how do we get the whole world connected? And Cisco was a big company
[13:40] connected? And Cisco was a big company
[13:40] connected? And Cisco was a big company behind this in kind of the late '9s,
[13:43] behind this in kind of the late '9s,
[13:43] behind this in kind of the late '9s, early 2000s. Uh, and uh, you'll hear
[13:46] early 2000s. Uh, and uh, you'll hear
[13:46] early 2000s. Uh, and uh, you'll hear this reference sometimes, but they they
[13:48] this reference sometimes, but they they
[13:48] this reference sometimes, but they they overbuilt. They were too far ahead of
[13:51] overbuilt. They were too far ahead of
[13:51] overbuilt. They were too far ahead of where the demand would be. And Cisco
[13:54] where the demand would be. And Cisco
[13:54] where the demand would be. And Cisco ended up paying for it. They, you know,
[13:56] ended up paying for it. They, you know,
[13:56] ended up paying for it. They, you know, they were one of the top stocks uh, of
[13:58] they were one of the top stocks uh, of
[13:58] they were one of the top stocks uh, of of the day and um, they peaked in, you
[14:02] of the day and um, they peaked in, you
[14:02] of the day and um, they peaked in, you know, early 2000s and it took them
[14:04] know, early 2000s and it took them
[14:04] know, early 2000s and it took them nearly 20 25 years to recover that same
[14:07] nearly 20 25 years to recover that same
[14:07] nearly 20 25 years to recover that same stock price. So this is not without risk
[14:09] stock price. So this is not without risk
[14:10] stock price. So this is not without risk of course. Um that being said, it looks
[14:12] of course. Um that being said, it looks
[14:12] of course. Um that being said, it looks like the demand is there and uh wish
[14:14] like the demand is there and uh wish
[14:14] like the demand is there and uh wish best of luck to Elon uh and the SpaceX
[14:18] best of luck to Elon uh and the SpaceX
[14:18] best of luck to Elon uh and the SpaceX AI team and looking forward to more
[14:20] AI team and looking forward to more
[14:20] AI team and looking forward to more compute, more power, more intelligence
[14:22] compute, more power, more intelligence
[14:22] compute, more power, more intelligence coming out of these models.
[14:25] coming out of these models.
[14:25] coming out of these models. Meta, Walmart, Stripe, and Sierra
[14:28] Meta, Walmart, Stripe, and Sierra
[14:28] Meta, Walmart, Stripe, and Sierra published an open standard for how AI
[14:30] published an open standard for how AI
[14:30] published an open standard for how AI agents shop and act on your behalf at
[14:34] agents shop and act on your behalf at
[14:34] agents shop and act on your behalf at businesses online. It's called the
[14:36] businesses online. It's called the
[14:36] businesses online. It's called the personal agent protocol
[14:38] personal agent protocol
[14:38] personal agent protocol using OOTH. This is the sign-in
[14:40] using OOTH. This is the sign-in
[14:40] using OOTH. This is the sign-in technology uh behind you know login with
[14:43] technology uh behind you know login with
[14:43] technology uh behind you know login with Google. Uh you would grant an agent
[14:46] Google. Uh you would grant an agent
[14:46] Google. Uh you would grant an agent specific read or write access to an
[14:48] specific read or write access to an
[14:48] specific read or write access to an account. The business decides what
[14:51] account. The business decides what
[14:51] account. The business decides what agents can do and can see uh what they
[14:54] agents can do and can see uh what they
[14:54] agents can do and can see uh what they did. So there's a kind of a trace there.
[14:56] did. So there's a kind of a trace there.
[14:56] did. So there's a kind of a trace there. Shopify, Genesis, Instinct, and Rocket
[15:00] Shopify, Genesis, Instinct, and Rocket
[15:00] Shopify, Genesis, Instinct, and Rocket also signed on, and a V0.1 spec is due
[15:05] also signed on, and a V0.1 spec is due
[15:05] also signed on, and a V0.1 spec is due later this month. Sierra co-founder
[15:07] later this month. Sierra co-founder
[15:07] later this month. Sierra co-founder Brett Taylor, who also chairs Open AI,
[15:10] Brett Taylor, who also chairs Open AI,
[15:10] Brett Taylor, who also chairs Open AI, told CN uh CNBC, "It is kind of chaos
[15:14] told CN uh CNBC, "It is kind of chaos
[15:14] told CN uh CNBC, "It is kind of chaos until such a standard exists." And I
[15:17] until such a standard exists." And I
[15:17] until such a standard exists." And I think he's he's right there. Right now,
[15:20] think he's he's right there. Right now,
[15:20] think he's he's right there. Right now, a store can't tell whether it's dealing
[15:21] a store can't tell whether it's dealing
[15:22] a store can't tell whether it's dealing with you or a bot acting for you.
[15:25] with you or a bot acting for you.
[15:25] with you or a bot acting for you. Whoever writes this standard shapes how
[15:27] Whoever writes this standard shapes how
[15:27] Whoever writes this standard shapes how AI shopping works. The backers say it
[15:31] AI shopping works. The backers say it
[15:31] AI shopping works. The backers say it lets a business verify that a bot is
[15:33] lets a business verify that a bot is
[15:33] lets a business verify that a bot is acting for a real person, which is
[15:34] acting for a real person, which is
[15:34] acting for a real person, which is great. Sierra published it Tuesday with
[15:37] great. Sierra published it Tuesday with
[15:37] great. Sierra published it Tuesday with the partners named and Taylor explained
[15:39] the partners named and Taylor explained
[15:39] the partners named and Taylor explained it to CNBC. It's on paper only. The spec
[15:42] it to CNBC. It's on paper only. The spec
[15:42] it to CNBC. It's on paper only. The spec isn't out yet, and Google, Amazon,
[15:44] isn't out yet, and Google, Amazon,
[15:44] isn't out yet, and Google, Amazon, Apple, and OpenAI aren't named partners.
[15:48] Apple, and OpenAI aren't named partners.
[15:48] Apple, and OpenAI aren't named partners. So, what does this this mean moving
[15:50] So, what does this this mean moving
[15:50] So, what does this this mean moving forward? um and and what are the
[15:53] forward? um and and what are the
[15:54] forward? um and and what are the implications and the context here. It's
[15:56] implications and the context here. It's
[15:56] implications and the context here. It's interesting. This is uh awesome to see
[15:58] interesting. This is uh awesome to see
[15:58] interesting. This is uh awesome to see this come out to see these folks working
[16:00] this come out to see these folks working
[16:00] this come out to see these folks working together. If you're not familiar, Sierra
[16:02] together. If you're not familiar, Sierra
[16:02] together. If you're not familiar, Sierra is a customer service uh agent company
[16:06] is a customer service uh agent company
[16:06] is a customer service uh agent company uh that has blown up. They are um
[16:09] uh that has blown up. They are um
[16:09] uh that has blown up. They are um they've grown very fast and and u from
[16:12] they've grown very fast and and u from
[16:12] they've grown very fast and and u from what I've heard uh their their product
[16:15] what I've heard uh their their product
[16:15] what I've heard uh their their product and the experience and the the agents
[16:17] and the experience and the the agents
[16:17] and the experience and the the agents that they offer, it's like magic, right?
[16:19] that they offer, it's like magic, right?
[16:19] that they offer, it's like magic, right? And so they are creating very very
[16:22] And so they are creating very very
[16:22] And so they are creating very very amazing experiences for their customers
[16:24] amazing experiences for their customers
[16:24] amazing experiences for their customers and transforming businesses. And this is
[16:26] and transforming businesses. And this is
[16:26] and transforming businesses. And this is kind of the next iteration possibly of
[16:29] kind of the next iteration possibly of
[16:29] kind of the next iteration possibly of what this would be. So instead of you
[16:31] what this would be. So instead of you
[16:31] what this would be. So instead of you know customer service agents you know
[16:33] know customer service agents you know
[16:33] know customer service agents you know handling direct customer issues as we
[16:36] handling direct customer issues as we
[16:36] handling direct customer issues as we see these personal agents come out again
[16:39] see these personal agents come out again
[16:39] see these personal agents come out again Grockbot uh OpenAI's dot Facebook or
[16:42] Grockbot uh OpenAI's dot Facebook or
[16:42] Grockbot uh OpenAI's dot Facebook or Meta's Muse um these agents want to go
[16:47] Meta's Muse um these agents want to go
[16:47] Meta's Muse um these agents want to go and act on behalf of humans and one of
[16:49] and act on behalf of humans and one of
[16:49] and act on behalf of humans and one of the biggest use cases that we're seeing
[16:50] the biggest use cases that we're seeing
[16:50] the biggest use cases that we're seeing is hey I need you to go you know buy
[16:53] is hey I need you to go you know buy
[16:53] is hey I need you to go you know buy groceries this week or I need you to go
[16:56] groceries this week or I need you to go
[16:56] groceries this week or I need you to go uh buy a a a gift for my mother-in-law.
[16:59] uh buy a a a gift for my mother-in-law.
[16:59] uh buy a a a gift for my mother-in-law. uh you know there there are different
[17:01] uh you know there there are different
[17:02] uh you know there there are different use cases like I don't have the time or
[17:03] use cases like I don't have the time or
[17:04] use cases like I don't have the time or don't want to go do that. Can I get my
[17:05] don't want to go do that. Can I get my
[17:05] don't want to go do that. Can I get my personal agent to go do it? Uh the
[17:08] personal agent to go do it? Uh the
[17:08] personal agent to go do it? Uh the problem is when you have a uh an agent
[17:11] problem is when you have a uh an agent
[17:11] problem is when you have a uh an agent kind of clicking around on the internet
[17:13] kind of clicking around on the internet
[17:14] kind of clicking around on the internet uh that that the internet is not set up
[17:17] uh that that the internet is not set up
[17:17] uh that that the internet is not set up for that, right? The internet is
[17:18] for that, right? The internet is
[17:18] for that, right? The internet is actually set up to block that, right? We
[17:20] actually set up to block that, right? We
[17:20] actually set up to block that, right? We don't want bots crawling around. Uh
[17:23] don't want bots crawling around. Uh
[17:23] don't want bots crawling around. Uh that's why you see, you know, captas and
[17:26] that's why you see, you know, captas and
[17:26] that's why you see, you know, captas and uh click if you're not, you know, click
[17:28] uh click if you're not, you know, click
[17:28] uh click if you're not, you know, click if you're not a robot, whatever, that
[17:30] if you're not a robot, whatever, that
[17:30] if you're not a robot, whatever, that kind of stuff. And so, um what what this
[17:33] kind of stuff. And so, um what what this
[17:33] kind of stuff. And so, um what what this sets up here is an actual standard and a
[17:36] sets up here is an actual standard and a
[17:36] sets up here is an actual standard and a protocol that says, "Hey, um you know,
[17:40] protocol that says, "Hey, um you know,
[17:40] protocol that says, "Hey, um you know, Walmart, I I want to go buy something.
[17:43] Walmart, I I want to go buy something.
[17:43] Walmart, I I want to go buy something. My agent wants to go buy something from
[17:45] My agent wants to go buy something from
[17:45] My agent wants to go buy something from you. This is my agent, and I've approved
[17:47] you. This is my agent, and I've approved
[17:47] you. This is my agent, and I've approved them to go act on my behalf." Walmart
[17:50] them to go act on my behalf." Walmart
[17:50] them to go act on my behalf." Walmart goes, "Okay, we recognize you as an
[17:52] goes, "Okay, we recognize you as an
[17:52] goes, "Okay, we recognize you as an agent. Agent come in and you can shop on
[17:54] agent. Agent come in and you can shop on
[17:54] agent. Agent come in and you can shop on behalf of Austin." Right? So, that
[17:57] behalf of Austin." Right? So, that
[17:57] behalf of Austin." Right? So, that that's the idea here. Um, one of the
[18:00] that's the idea here. Um, one of the
[18:00] that's the idea here. Um, one of the biggest problems, we've seen this uh in
[18:02] biggest problems, we've seen this uh in
[18:02] biggest problems, we've seen this uh in a couple different stories come out from
[18:04] a couple different stories come out from
[18:04] a couple different stories come out from Amazon and that they are actively
[18:07] Amazon and that they are actively
[18:07] Amazon and that they are actively blocking agents from purchasing on uh on
[18:11] blocking agents from purchasing on uh on
[18:11] blocking agents from purchasing on uh on the the Amazon web app or on the um
[18:15] the the Amazon web app or on the um
[18:15] the the Amazon web app or on the um yeah, the Amazon store in general. The
[18:18] yeah, the Amazon store in general. The
[18:18] yeah, the Amazon store in general. The big problem there is Amazon uh generates
[18:21] big problem there is Amazon uh generates
[18:21] big problem there is Amazon uh generates a great deal of their revenue from um ad
[18:24] a great deal of their revenue from um ad
[18:24] a great deal of their revenue from um ad spend. And so they have advertisers that
[18:27] spend. And so they have advertisers that
[18:27] spend. And so they have advertisers that are and you you barely notice it, but
[18:29] are and you you barely notice it, but
[18:29] are and you you barely notice it, but you know when you search for uh hey, I
[18:32] you know when you search for uh hey, I
[18:32] you know when you search for uh hey, I need to buy some uh light bulbs, you're
[18:34] need to buy some uh light bulbs, you're
[18:34] need to buy some uh light bulbs, you're going to see uh a sponsored search pop
[18:37] going to see uh a sponsored search pop
[18:37] going to see uh a sponsored search pop up. if you have an agent doing that
[18:40] up. if you have an agent doing that
[18:40] up. if you have an agent doing that search, those uh searches become less
[18:43] search, those uh searches become less
[18:43] search, those uh searches become less and less valuable to Amazon and to their
[18:46] and less valuable to Amazon and to their
[18:46] and less valuable to Amazon and to their advertisers. And so there's a kind of a
[18:49] advertisers. And so there's a kind of a
[18:49] advertisers. And so there's a kind of a complex issue here. It'll be interesting
[18:51] complex issue here. It'll be interesting
[18:51] complex issue here. It'll be interesting to see how Amazon and you know, you've
[18:54] to see how Amazon and you know, you've
[18:54] to see how Amazon and you know, you've got to imagine it. It says Amazon and
[18:56] got to imagine it. It says Amazon and
[18:56] got to imagine it. It says Amazon and and Apple aren't on board. um you know,
[19:00] and Apple aren't on board. um you know,
[19:00] and Apple aren't on board. um you know, is Apple going to want to let agents uh
[19:05] is Apple going to want to let agents uh
[19:05] is Apple going to want to let agents uh kind of crawl around their ecosystem and
[19:07] kind of crawl around their ecosystem and
[19:07] kind of crawl around their ecosystem and make decisions and purchase stuff as
[19:09] make decisions and purchase stuff as
[19:10] make decisions and purchase stuff as well too? So that it'll be interesting
[19:11] well too? So that it'll be interesting
[19:11] well too? So that it'll be interesting to see if all of the u technology
[19:15] to see if all of the u technology
[19:15] to see if all of the u technology companies come together with the uh
[19:18] companies come together with the uh
[19:18] companies come together with the uh consumer brands or retail companies and
[19:20] consumer brands or retail companies and
[19:20] consumer brands or retail companies and how they work together. And if we do
[19:22] how they work together. And if we do
[19:22] how they work together. And if we do come up with a single agent protocol, I
[19:25] come up with a single agent protocol, I
[19:26] come up with a single agent protocol, I kind of can't imagine that that we come
[19:28] kind of can't imagine that that we come
[19:28] kind of can't imagine that that we come up with one protocol that everybody
[19:29] up with one protocol that everybody
[19:30] up with one protocol that everybody agrees on, but it'll be interesting to
[19:31] agrees on, but it'll be interesting to
[19:31] agrees on, but it'll be interesting to see if that's the the case. That being
[19:33] see if that's the the case. That being
[19:33] see if that's the the case. That being said, um we're we're moving forward and
[19:37] said, um we're we're moving forward and
[19:37] said, um we're we're moving forward and um and and the the direct impact to you
[19:40] um and and the the direct impact to you
[19:40] um and and the the direct impact to you is uh now you you can obiscate some of
[19:44] is uh now you you can obiscate some of
[19:44] is uh now you you can obiscate some of the time that you spent uh shopping. you
[19:48] the time that you spent uh shopping. you
[19:48] the time that you spent uh shopping. you know, my wife is shopping for a new sofa
[19:51] know, my wife is shopping for a new sofa
[19:51] know, my wife is shopping for a new sofa and it's just can't find the right one,
[19:53] and it's just can't find the right one,
[19:53] and it's just can't find the right one, can't do it. Hey, can I have my agent go
[19:55] can't do it. Hey, can I have my agent go
[19:55] can't do it. Hey, can I have my agent go find me a few options and come back and
[19:57] find me a few options and come back and
[19:57] find me a few options and come back and let me know? And then can I tell my
[19:58] let me know? And then can I tell my
[19:58] let me know? And then can I tell my agent, hey, go ahead and purchase that
[20:00] agent, hey, go ahead and purchase that
[20:00] agent, hey, go ahead and purchase that for me. Um the the amount of time that
[20:03] for me. Um the the amount of time that
[20:03] for me. Um the the amount of time that we'll save here is amazing. And then the
[20:05] we'll save here is amazing. And then the
[20:05] we'll save here is amazing. And then the interesting, you know, related impact
[20:07] interesting, you know, related impact
[20:07] interesting, you know, related impact that we see is on the economy. Like do
[20:10] that we see is on the economy. Like do
[20:10] that we see is on the economy. Like do we see retail and consumer spending
[20:12] we see retail and consumer spending
[20:12] we see retail and consumer spending increase by any measurable uh number
[20:15] increase by any measurable uh number
[20:15] increase by any measurable uh number because of these agents? Um, we'll
[20:17] because of these agents? Um, we'll
[20:17] because of these agents? Um, we'll continue to watch this closely, but
[20:19] continue to watch this closely, but
[20:19] continue to watch this closely, but again, great to see the the teams
[20:20] again, great to see the the teams
[20:20] again, great to see the the teams working together and seeing some
[20:22] working together and seeing some
[20:22] working together and seeing some advancement here. France's Mistral
[20:25] advancement here. France's Mistral
[20:25] advancement here. France's Mistral unveiled large for nicknamed Lechon, a
[20:29] unveiled large for nicknamed Lechon, a
[20:29] unveiled large for nicknamed Lechon, a model of about one trillion parameters
[20:31] model of about one trillion parameters
[20:32] model of about one trillion parameters that it calls the most powerful
[20:33] that it calls the most powerful
[20:33] that it calls the most powerful openweight AI outside of uh, China.
[20:38] openweight AI outside of uh, China.
[20:38] openweight AI outside of uh, China. It has uh, a little over one trillion
[20:40] It has uh, a little over one trillion
[20:40] It has uh, a little over one trillion parameters with 49 billion active per
[20:44] parameters with 49 billion active per
[20:44] parameters with 49 billion active per word. the same mixture of experts design
[20:47] word. the same mixture of experts design
[20:47] word. the same mixture of experts design as reflections beam which we covered
[20:49] as reflections beam which we covered
[20:49] as reflections beam which we covered yesterday. Uh Mistrol trained it from
[20:51] yesterday. Uh Mistrol trained it from
[20:52] yesterday. Uh Mistrol trained it from scratch on 4,000 Nvidia Grace Blackwell
[20:55] scratch on 4,000 Nvidia Grace Blackwell
[20:55] scratch on 4,000 Nvidia Grace Blackwell GPUs over 2 months in its own European
[20:59] GPUs over 2 months in its own European
[20:59] GPUs over 2 months in its own European data centers and it covers more than 160
[21:01] data centers and it covers more than 160
[21:01] data centers and it covers more than 160 languages. Developers can use it through
[21:04] languages. Developers can use it through
[21:04] languages. Developers can use it through Mistl's API now. Uh the weights are due
[21:08] Mistl's API now. Uh the weights are due
[21:08] Mistl's API now. Uh the weights are due October 27th.
[21:11] October 27th.
[21:11] October 27th. two days, two big non-Chinese open
[21:14] two days, two big non-Chinese open
[21:14] two days, two big non-Chinese open models dropping. The openweight race is
[21:16] models dropping. The openweight race is
[21:16] models dropping. The openweight race is no longer China's alone. Uh Misl says
[21:20] no longer China's alone. Uh Misl says
[21:20] no longer China's alone. Uh Misl says it's the most powerful openweight AI
[21:22] it's the most powerful openweight AI
[21:22] it's the most powerful openweight AI system outside of China. Co-founder Giam
[21:25] system outside of China. Co-founder Giam
[21:25] system outside of China. Co-founder Giam Lampo says ML4 is at the frontier of
[21:29] Lampo says ML4 is at the frontier of
[21:29] Lampo says ML4 is at the frontier of openweight models. ML's own benchmarks
[21:32] openweight models. ML's own benchmarks
[21:32] openweight models. ML's own benchmarks uh say that it scores 62% on the
[21:37] uh say that it scores 62% on the
[21:37] uh say that it scores 62% on the software engineer focused deepu. Uh,
[21:40] software engineer focused deepu. Uh,
[21:40] software engineer focused deepu. Uh, VentureB notes that score doesn't
[21:42] VentureB notes that score doesn't
[21:42] VentureB notes that score doesn't establish an outright coding lead. Uh, I
[21:44] establish an outright coding lead. Uh, I
[21:44] establish an outright coding lead. Uh, I don't think um, but for open sources,
[21:46] don't think um, but for open sources,
[21:46] don't think um, but for open sources, that's a great score, right? Uh, you
[21:48] that's a great score, right? Uh, you
[21:48] that's a great score, right? Uh, you can't download it until October 27th and
[21:51] can't download it until October 27th and
[21:51] can't download it until October 27th and it ships under Mistl's own license, not
[21:54] it ships under Mistl's own license, not
[21:54] it ships under Mistl's own license, not a standard open source one, which uh is
[21:56] a standard open source one, which uh is
[21:56] a standard open source one, which uh is an interesting uh we'll we'll see what
[21:58] an interesting uh we'll we'll see what
[21:58] an interesting uh we'll we'll see what that actually means. So, uh, again,
[22:01] that actually means. So, uh, again,
[22:01] that actually means. So, uh, again, seeing progress here, seeing progress
[22:03] seeing progress here, seeing progress
[22:03] seeing progress here, seeing progress now out of Europe, which is wonderful.
[22:04] now out of Europe, which is wonderful.
[22:04] now out of Europe, which is wonderful. We love competition. We love more model
[22:07] We love competition. We love more model
[22:08] We love competition. We love more model companies in the game and we love more
[22:09] companies in the game and we love more
[22:09] companies in the game and we love more open source model companies in the game.
[22:11] open source model companies in the game.
[22:11] open source model companies in the game. It's interesting. So in in this story we
[22:13] It's interesting. So in in this story we
[22:13] It's interesting. So in in this story we we we reference Mistral training on
[22:16] we we reference Mistral training on
[22:16] we we reference Mistral training on 4,000 Nvidia Grace Blackwell GPUs. So
[22:20] 4,000 Nvidia Grace Blackwell GPUs. So
[22:20] 4,000 Nvidia Grace Blackwell GPUs. So those are the chips uh 4,000 of those uh
[22:24] those are the chips uh 4,000 of those uh
[22:24] those are the chips uh 4,000 of those uh that they trained on it for several
[22:25] that they trained on it for several
[22:25] that they trained on it for several months to generate this model. It has a
[22:30] months to generate this model. It has a
[22:30] months to generate this model. It has a trillion parameters. That means that uh
[22:33] trillion parameters. That means that uh
[22:33] trillion parameters. That means that uh basically it has a trillion little rules
[22:35] basically it has a trillion little rules
[22:35] basically it has a trillion little rules in it that each time in a standard LLM
[22:39] in it that each time in a standard LLM
[22:39] in it that each time in a standard LLM each time that you ask a question it's
[22:42] each time that you ask a question it's
[22:42] each time that you ask a question it's going to run through all one trillion of
[22:44] going to run through all one trillion of
[22:44] going to run through all one trillion of those. In this case, just like with the
[22:47] those. In this case, just like with the
[22:47] those. In this case, just like with the the model that we covered yesterday, uh
[22:50] the model that we covered yesterday, uh
[22:50] the model that we covered yesterday, uh the reflection beam model, um when you
[22:53] the reflection beam model, um when you
[22:53] the reflection beam model, um when you ask a question, instead of running it
[22:54] ask a question, instead of running it
[22:54] ask a question, instead of running it through a trillion parameters, it just
[22:57] through a trillion parameters, it just
[22:57] through a trillion parameters, it just cuts a little slice, 49 billion
[23:00] cuts a little slice, 49 billion
[23:00] cuts a little slice, 49 billion parameters, and it runs it through
[23:01] parameters, and it runs it through
[23:02] parameters, and it runs it through those. Uh that's why we need the
[23:04] those. Uh that's why we need the
[23:04] those. Uh that's why we need the compute. That's why we need the data
[23:05] compute. That's why we need the data
[23:06] compute. That's why we need the data centers. That's why Elon's getting a $40
[23:08] centers. That's why Elon's getting a $40
[23:08] centers. That's why Elon's getting a $40 billion loan. Um, and that's why we
[23:10] billion loan. Um, and that's why we
[23:10] billion loan. Um, and that's why we continue to see progress in the hardware
[23:13] continue to see progress in the hardware
[23:13] continue to see progress in the hardware space because these models and this
[23:16] space because these models and this
[23:16] space because these models and this intelligence and pushing the frontier of
[23:18] intelligence and pushing the frontier of
[23:18] intelligence and pushing the frontier of this intelligent is so compute heavy. It
[23:21] this intelligent is so compute heavy. It
[23:21] this intelligent is so compute heavy. It requires, you know, in this case 4,000
[23:25] requires, you know, in this case 4,000
[23:25] requires, you know, in this case 4,000 uh Grace Blackwell chips running for 2
[23:27] uh Grace Blackwell chips running for 2
[23:28] uh Grace Blackwell chips running for 2 months constantly, right? It's just
[23:30] months constantly, right? It's just
[23:30] months constantly, right? It's just constantly running. And then when we get
[23:32] constantly running. And then when we get
[23:32] constantly running. And then when we get users on the other end, they're going to
[23:34] users on the other end, they're going to
[23:34] users on the other end, they're going to be using uh the the the chips as well to
[23:36] be using uh the the the chips as well to
[23:36] be using uh the the the chips as well to to actually run the mo models on them
[23:38] to actually run the mo models on them
[23:38] to actually run the mo models on them and and use them and ask questions and
[23:41] and and use them and ask questions and
[23:41] and and use them and ask questions and and uncover, you know, 372 math
[23:44] and uncover, you know, 372 math
[23:44] and uncover, you know, 372 math problems, right? Not [snorts] to uh
[23:46] problems, right? Not [snorts] to uh
[23:46] problems, right? Not [snorts] to uh overlap too many stories there, but this
[23:48] overlap too many stories there, but this
[23:48] overlap too many stories there, but this is what's happening. Um all in all,
[23:51] is what's happening. Um all in all,
[23:51] is what's happening. Um all in all, congrats to the team at Mistral. Uh,
[23:55] congrats to the team at Mistral. Uh,
[23:55] congrats to the team at Mistral. Uh, congrats to uh Europe and and to France
[23:58] congrats to uh Europe and and to France
[23:58] congrats to uh Europe and and to France and and like we want to see this
[24:01] and and like we want to see this
[24:01] and and like we want to see this everywhere. We want to see everybody
[24:03] everywhere. We want to see everybody
[24:03] everywhere. We want to see everybody coming out with new models um and new
[24:07] coming out with new models um and new
[24:07] coming out with new models um and new open source new openweight models which
[24:09] open source new openweight models which
[24:09] open source new openweight models which means that anybody can use these um
[24:11] means that anybody can use these um
[24:11] means that anybody can use these um they're they're not restricted um you
[24:14] they're they're not restricted um you
[24:14] they're they're not restricted um you know and and especially from a cost
[24:16] know and and especially from a cost
[24:16] know and and especially from a cost standpoint or from a compute standpoint
[24:19] standpoint or from a compute standpoint
[24:19] standpoint or from a compute standpoint this opens up the availability for
[24:20] this opens up the availability for
[24:20] this opens up the availability for people to use this across the world. Um,
[24:23] people to use this across the world. Um,
[24:23] people to use this across the world. Um, and looks like we're going to get that
[24:25] and looks like we're going to get that
[24:25] and looks like we're going to get that uh towards the end of this month. And
[24:26] uh towards the end of this month. And
[24:26] uh towards the end of this month. And same same thing with the uh reflection
[24:28] same same thing with the uh reflection
[24:28] same same thing with the uh reflection model, which is really cool. So,
[24:30] model, which is really cool. So,
[24:30] model, which is really cool. So, congrats to that team over there.
[24:32] congrats to that team over there.
[24:32] congrats to that team over there. Excited for them. Uh, excited to see
[24:34] Excited for them. Uh, excited to see
[24:34] Excited for them. Uh, excited to see more come out of Europe. Um, and to see
[24:37] more come out of Europe. Um, and to see
[24:37] more come out of Europe. Um, and to see more advancements in open source. The AI
[24:40] more advancements in open source. The AI
[24:40] more advancements in open source. The AI cloud company Lambda is reportedly
[24:42] cloud company Lambda is reportedly
[24:42] cloud company Lambda is reportedly raising up to $4 billion at a 14.5
[24:45] raising up to $4 billion at a 14.5
[24:45] raising up to $4 billion at a 14.5 billion valuation, its last private
[24:48] billion valuation, its last private
[24:48] billion valuation, its last private round before they go public. The round
[24:50] round before they go public. The round
[24:50] round before they go public. The round of up to $4 billion values Lambda at
[24:53] of up to $4 billion values Lambda at
[24:53] of up to $4 billion values Lambda at $14.5 billion uh before the new money.
[24:57] $14.5 billion uh before the new money.
[24:57] $14.5 billion uh before the new money. Blackstone and CO2 are leading it, the
[25:00] Blackstone and CO2 are leading it, the
[25:00] Blackstone and CO2 are leading it, the Wall Street Journal reported Tuesday.
[25:03] Wall Street Journal reported Tuesday.
[25:03] Wall Street Journal reported Tuesday. And Lambda is aiming for a 2027 IPO.
[25:07] And Lambda is aiming for a 2027 IPO.
[25:07] And Lambda is aiming for a 2027 IPO. Its order backlog jumped from $15
[25:10] Its order backlog jumped from $15
[25:10] Its order backlog jumped from $15 billion in June to $50 billion in
[25:13] billion in June to $50 billion in
[25:13] billion in June to $50 billion in September, anchored by a $35 billion
[25:15] September, anchored by a $35 billion
[25:15] September, anchored by a $35 billion compute deal with Anthropic. Last
[25:18] compute deal with Anthropic. Last
[25:18] compute deal with Anthropic. Last November it was valued at uh $5.9
[25:22] November it was valued at uh $5.9
[25:22] November it was valued at uh $5.9 billion and this is just crazy to see
[25:25] billion and this is just crazy to see
[25:25] billion and this is just crazy to see over and over again. So this pigs and
[25:27] over and over again. So this pigs and
[25:27] over and over again. So this pigs and shovels trade the companies uh basically
[25:31] shovels trade the companies uh basically
[25:31] shovels trade the companies uh basically in in this case Lambda rents Nvidia
[25:34] in in this case Lambda rents Nvidia
[25:34] in in this case Lambda rents Nvidia chips to AI labs. So to the mistrals of
[25:38] chips to AI labs. So to the mistrals of
[25:38] chips to AI labs. So to the mistrals of the world to the reflection beams of the
[25:41] the world to the reflection beams of the
[25:41] the world to the reflection beams of the world to um the open AI to anthropic
[25:45] world to um the open AI to anthropic
[25:45] world to um the open AI to anthropic here right uh so the companies renting
[25:48] here right uh so the companies renting
[25:48] here right uh so the companies renting Nvidia chips to AI labs are growing as
[25:50] Nvidia chips to AI labs are growing as
[25:50] Nvidia chips to AI labs are growing as fast as the AI labs themselves. Lambda
[25:53] fast as the AI labs themselves. Lambda
[25:53] fast as the AI labs themselves. Lambda is one of the quote unquote neoclouds
[25:56] is one of the quote unquote neoclouds
[25:56] is one of the quote unquote neoclouds that rent Nvidia GPUs to AI labs and
[25:59] that rent Nvidia GPUs to AI labs and
[25:59] that rent Nvidia GPUs to AI labs and Nvidia is an investor. So this is good
[26:01] Nvidia is an investor. So this is good
[26:01] Nvidia is an investor. So this is good for Nvidia. the $50 billion backlog and
[26:05] for Nvidia. the $50 billion backlog and
[26:05] for Nvidia. the $50 billion backlog and the anthropic deal uh are what's really
[26:09] the anthropic deal uh are what's really
[26:09] the anthropic deal uh are what's really driving uh this growth and this new
[26:11] driving uh this growth and this new
[26:11] driving uh this growth and this new funding round. Uh a backlog is promised
[26:15] funding round. Uh a backlog is promised
[26:15] funding round. Uh a backlog is promised future spending and one customer
[26:17] future spending and one customer
[26:17] future spending and one customer anthropic accounts for most of it which
[26:19] anthropic accounts for most of it which
[26:19] anthropic accounts for most of it which does create some risk there but uh
[26:22] does create some risk there but uh
[26:22] does create some risk there but uh justifies the additional spend and and
[26:26] justifies the additional spend and and
[26:26] justifies the additional spend and and growth needed. Uh again this kind of
[26:28] growth needed. Uh again this kind of
[26:28] growth needed. Uh again this kind of comes back to the question you know are
[26:31] comes back to the question you know are
[26:31] comes back to the question you know are we getting overbuilt? It doesn't seem
[26:33] we getting overbuilt? It doesn't seem
[26:33] we getting overbuilt? It doesn't seem like it at this point. This is the big
[26:35] like it at this point. This is the big
[26:35] like it at this point. This is the big difference between kind of the internet
[26:37] difference between kind of the internet
[26:37] difference between kind of the internet age and where we are at currently. Uh in
[26:40] age and where we are at currently. Uh in
[26:40] age and where we are at currently. Uh in that um in the internet pets.com didn't
[26:43] that um in the internet pets.com didn't
[26:43] that um in the internet pets.com didn't have any revenue and uh little dotcoms
[26:47] have any revenue and uh little dotcoms
[26:47] have any revenue and uh little dotcoms were going public at you know tens of
[26:49] were going public at you know tens of
[26:49] were going public at you know tens of million dollar valuations whatever but
[26:52] million dollar valuations whatever but
[26:52] million dollar valuations whatever but didn't have revenue or had negative re
[26:54] didn't have revenue or had negative re
[26:54] didn't have revenue or had negative re revenue whatever the case may be. In
[26:56] revenue whatever the case may be. In
[26:56] revenue whatever the case may be. In this case, we see Lambda has $50 billion
[27:00] this case, we see Lambda has $50 billion
[27:00] this case, we see Lambda has $50 billion backlog. It seems like most of um this
[27:04] backlog. It seems like most of um this
[27:04] backlog. It seems like most of um this uh commit uh compute commit is coming
[27:07] uh commit uh compute commit is coming
[27:07] uh commit uh compute commit is coming from Anthropic. But needless to say um
[27:10] from Anthropic. But needless to say um
[27:10] from Anthropic. But needless to say um Anthropic is about to go public and open
[27:13] Anthropic is about to go public and open
[27:13] Anthropic is about to go public and open up a bunch of free cash where they will
[27:16] up a bunch of free cash where they will
[27:16] up a bunch of free cash where they will continue to spend on uh compute. Uh you
[27:20] continue to spend on uh compute. Uh you
[27:20] continue to spend on uh compute. Uh you know, we see this constraint. Open AAI,
[27:22] know, we see this constraint. Open AAI,
[27:22] know, we see this constraint. Open AAI, we've covered a couple things around
[27:23] we've covered a couple things around
[27:23] we've covered a couple things around this where OpenAI has done some things
[27:25] this where OpenAI has done some things
[27:25] this where OpenAI has done some things in the past uh week or so that would
[27:28] in the past uh week or so that would
[27:28] in the past uh week or so that would signal maybe they don't have uh compute
[27:31] signal maybe they don't have uh compute
[27:31] signal maybe they don't have uh compute capacity. Um I brought this up to
[27:33] capacity. Um I brought this up to
[27:33] capacity. Um I brought this up to somebody this week and they said it's
[27:35] somebody this week and they said it's
[27:35] somebody this week and they said it's it's not a compute capacity problem,
[27:36] it's not a compute capacity problem,
[27:36] it's not a compute capacity problem, it's a money problem. They don't they
[27:38] it's a money problem. They don't they
[27:38] it's a money problem. They don't they don't have enough money to pay for the
[27:40] don't have enough money to pay for the
[27:40] don't have enough money to pay for the compute. And so the demand is there.
[27:42] compute. And so the demand is there.
[27:42] compute. And so the demand is there. It's all about getting the money to flow
[27:44] It's all about getting the money to flow
[27:44] It's all about getting the money to flow in the right direction. uh the revenue
[27:47] in the right direction. uh the revenue
[27:47] in the right direction. uh the revenue is there for these um you know neoclouds
[27:51] is there for these um you know neoclouds
[27:51] is there for these um you know neoclouds and for the folks at the bottom of the
[27:53] and for the folks at the bottom of the
[27:53] and for the folks at the bottom of the stack right the infrastructure layer uh
[27:56] stack right the infrastructure layer uh
[27:56] stack right the infrastructure layer uh the chip companies we talked about
[27:57] the chip companies we talked about
[27:57] the chip companies we talked about etched yesterday um and so there's
[28:01] etched yesterday um and so there's
[28:01] etched yesterday um and so there's definitely demand here uh whether or not
[28:04] definitely demand here uh whether or not
[28:04] definitely demand here uh whether or not it's oversized I think we're so early
[28:08] it's oversized I think we're so early
[28:08] it's oversized I think we're so early you know we talked about uh only 2% or
[28:11] you know we talked about uh only 2% or
[28:11] you know we talked about uh only 2% or something of consumers pay for uh AI
[28:15] something of consumers pay for uh AI
[28:15] something of consumers pay for uh AI services, man. Um, we'll be able to see
[28:19] services, man. Um, we'll be able to see
[28:19] services, man. Um, we'll be able to see that grow substantially
[28:22] that grow substantially
[28:22] that grow substantially uh over the next year. And I think
[28:24] uh over the next year. And I think
[28:24] uh over the next year. And I think especially in my estimation as we see
[28:27] especially in my estimation as we see
[28:27] especially in my estimation as we see these breakthroughs from open AI uh as
[28:30] these breakthroughs from open AI uh as
[28:30] these breakthroughs from open AI uh as we continue to, you know, see open
[28:32] we continue to, you know, see open
[28:32] we continue to, you know, see open source grow, we we see the pace of of
[28:35] source grow, we we see the pace of of
[28:35] source grow, we we see the pace of of development advancement speed up. I
[28:38] development advancement speed up. I
[28:38] development advancement speed up. I believe that by the end of 2027
[28:41] believe that by the end of 2027
[28:41] believe that by the end of 2027 um we're going to have substantial
[28:42] um we're going to have substantial
[28:42] um we're going to have substantial breakthrough such that um the everyday
[28:47] breakthrough such that um the everyday
[28:47] breakthrough such that um the everyday you know our our everyday lives will be
[28:49] you know our our everyday lives will be
[28:50] you know our our everyday lives will be dramatically impacted in a very positive
[28:52] dramatically impacted in a very positive
[28:52] dramatically impacted in a very positive way. uh and most of us will not look at
[28:55] way. uh and most of us will not look at
[28:55] way. uh and most of us will not look at the world in in the same way probably by
[28:58] the world in in the same way probably by
[28:58] the world in in the same way probably by the end of 2027. As you look forward to
[29:02] the end of 2027. As you look forward to
[29:02] the end of 2027. As you look forward to uh you know the next year 2028 you know
[29:06] uh you know the next year 2028 you know
[29:06] uh you know the next year 2028 you know you look at two years out and it's like
[29:07] you look at two years out and it's like
[29:07] you look at two years out and it's like man I you can only imagine
[29:10] man I you can only imagine
[29:10] man I you can only imagine [clears throat]
[29:10] [clears throat]
[29:10] [clears throat] what the world is going to look like.
[29:12] what the world is going to look like.
[29:12] what the world is going to look like. It's going to be a completely different
[29:14] It's going to be a completely different
[29:14] It's going to be a completely different place and I'm excited for it. I hope
[29:16] place and I'm excited for it. I hope
[29:16] place and I'm excited for it. I hope you're excited for it. We're excited for
[29:18] you're excited for it. We're excited for
[29:18] you're excited for it. We're excited for Lambda and uh uh wish them the best of
[29:21] Lambda and uh uh wish them the best of
[29:21] Lambda and uh uh wish them the best of luck on their funding round and
[29:25] luck on their funding round and
[29:25] luck on their funding round and developing more compute.
[29:27] developing more compute.
[29:27] developing more compute. Google released Nano Banana 2.1. This is
[29:30] Google released Nano Banana 2.1. This is
[29:30] Google released Nano Banana 2.1. This is an upgrade to its popular AI image
[29:33] an upgrade to its popular AI image
[29:33] an upgrade to its popular AI image generator and editor built on Gemini 3.6
[29:37] generator and editor built on Gemini 3.6
[29:37] generator and editor built on Gemini 3.6 Flash. Nanobanana 2.1 is built on uh
[29:40] Flash. Nanobanana 2.1 is built on uh
[29:40] Flash. Nanobanana 2.1 is built on uh Gemini 3.6 6 Flash and is rolling out in
[29:43] Gemini 3.6 6 Flash and is rolling out in
[29:43] Gemini 3.6 6 Flash and is rolling out in the Gemini app. Google AI Stud Studio
[29:46] the Gemini app. Google AI Stud Studio
[29:46] the Gemini app. Google AI Stud Studio and uh the Gemini API. Google says it
[29:49] and uh the Gemini API. Google says it
[29:49] and uh the Gemini API. Google says it improves on previous versions across the
[29:52] improves on previous versions across the
[29:52] improves on previous versions across the board. Developers pay about 3 cents for
[29:55] board. Developers pay about 3 cents for
[29:55] board. Developers pay about 3 cents for a standard 1K image and about 7.6 cents
[29:59] a standard 1K image and about 7.6 cents
[29:59] a standard 1K image and about 7.6 cents uh for a 4K image. We love to see more
[30:03] uh for a 4K image. We love to see more
[30:03] uh for a 4K image. We love to see more images. Image editing is one of the AI
[30:06] images. Image editing is one of the AI
[30:06] images. Image editing is one of the AI features regular people actually use.
[30:08] features regular people actually use.
[30:08] features regular people actually use. And this lands uh right in the Gemini
[30:10] And this lands uh right in the Gemini
[30:10] And this lands uh right in the Gemini app, which is really nice, especially
[30:12] app, which is really nice, especially
[30:12] app, which is really nice, especially coming on the heels of Gemini's Argon
[30:14] coming on the heels of Gemini's Argon
[30:14] coming on the heels of Gemini's Argon release last week. Google says it beat
[30:17] release last week. Google says it beat
[30:17] release last week. Google says it beat uh earlier versions across the board.
[30:20] uh earlier versions across the board.
[30:20] uh earlier versions across the board. Alpha Signal reports uh that the new
[30:23] Alpha Signal reports uh that the new
[30:23] Alpha Signal reports uh that the new Nano Banana beats the Pro model on
[30:25] Nano Banana beats the Pro model on
[30:25] Nano Banana beats the Pro model on editing. Google's release notes uh and
[30:30] editing. Google's release notes uh and
[30:30] editing. Google's release notes uh and published API prices uh are are where
[30:34] published API prices uh are are where
[30:34] published API prices uh are are where this story is coming from. Reports
[30:35] this story is coming from. Reports
[30:36] this story is coming from. Reports disagree on price. Uh the decoder says
[30:38] disagree on price. Uh the decoder says
[30:38] disagree on price. Uh the decoder says about half the cost of Nana Banana 2
[30:41] about half the cost of Nana Banana 2
[30:41] about half the cost of Nana Banana 2 while others say the price is unchanged.
[30:43] while others say the price is unchanged.
[30:43] while others say the price is unchanged. Uh and Google's claims are the only
[30:46] Uh and Google's claims are the only
[30:46] Uh and Google's claims are the only benchmark so for so far. So we'll have
[30:48] benchmark so for so far. So we'll have
[30:48] benchmark so for so far. So we'll have to see where where the cost lands on
[30:49] to see where where the cost lands on
[30:49] to see where where the cost lands on this. But um you know I was just
[30:53] this. But um you know I was just
[30:53] this. But um you know I was just thinking
[30:56] thinking
[30:56] thinking do we have enough AI slop out there? I
[30:58] do we have enough AI slop out there? I
[30:58] do we have enough AI slop out there? I don't think so. We need more slop. Uh
[31:01] don't think so. We need more slop. Uh
[31:02] don't think so. We need more slop. Uh sloptober.
[31:03] sloptober.
[31:03] sloptober. There's a uh uh dang what is the name of
[31:07] There's a uh uh dang what is the name of
[31:07] There's a uh uh dang what is the name of this? There's a company doing Sloptober
[31:11] this? There's a company doing Sloptober
[31:11] this? There's a company doing Sloptober um which is all AI generated uh videos
[31:15] um which is all AI generated uh videos
[31:15] um which is all AI generated uh videos and they're doing a whole contest around
[31:16] and they're doing a whole contest around
[31:16] and they're doing a whole contest around it. Mark Andre got behind it. We need
[31:19] it. Mark Andre got behind it. We need
[31:19] it. Mark Andre got behind it. We need more slot man. We need more AI images.
[31:22] more slot man. We need more AI images.
[31:22] more slot man. We need more AI images. Uh they are getting better and better.
[31:24] Uh they are getting better and better.
[31:24] Uh they are getting better and better. It is phenomenal to see uh what these
[31:27] It is phenomenal to see uh what these
[31:27] It is phenomenal to see uh what these models can do and and just the
[31:28] models can do and and just the
[31:28] models can do and and just the capability. I think it's really
[31:31] capability. I think it's really
[31:31] capability. I think it's really interesting. you know, not very long
[31:33] interesting. you know, not very long
[31:33] interesting. you know, not very long ago, it was like, uh, hey, generate this
[31:36] ago, it was like, uh, hey, generate this
[31:36] ago, it was like, uh, hey, generate this image, and it was not really very close
[31:39] image, and it was not really very close
[31:39] image, and it was not really very close to what you were asking for. Maybe it
[31:41] to what you were asking for. Maybe it
[31:41] to what you were asking for. Maybe it looked really good, but the the
[31:42] looked really good, but the the
[31:42] looked really good, but the the accuracy, let's say, was not that great.
[31:45] accuracy, let's say, was not that great.
[31:45] accuracy, let's say, was not that great. Um, that is getting much much better.
[31:47] Um, that is getting much much better.
[31:47] Um, that is getting much much better. And then it's going beyond just image
[31:49] And then it's going beyond just image
[31:49] And then it's going beyond just image and going to um to video, which is uh
[31:54] and going to um to video, which is uh
[31:54] and going to um to video, which is uh crazy to see when you can say, "Hey, I
[31:56] crazy to see when you can say, "Hey, I
[31:56] crazy to see when you can say, "Hey, I need a video uh you know, that does X,
[31:59] need a video uh you know, that does X,
[31:59] need a video uh you know, that does X, Y, and Z." And it goes, "Oh, you mean
[32:00] Y, and Z." And it goes, "Oh, you mean
[32:00] Y, and Z." And it goes, "Oh, you mean this?" and you're like, "Yeah, that's
[32:01] this?" and you're like, "Yeah, that's
[32:02] this?" and you're like, "Yeah, that's exactly what I wanted." Um uh and maybe
[32:04] exactly what I wanted." Um uh and maybe
[32:04] exactly what I wanted." Um uh and maybe there's like a few tweaks, you know, you
[32:06] there's like a few tweaks, you know, you
[32:06] there's like a few tweaks, you know, you you do a couple different turns back and
[32:08] you do a couple different turns back and
[32:08] you do a couple different turns back and forth, but altogether it's um they're
[32:11] forth, but altogether it's um they're
[32:11] forth, but altogether it's um they're they're turning out really really good
[32:14] they're turning out really really good
[32:14] they're turning out really really good content. So, good to see uh Nano Banana.
[32:17] content. So, good to see uh Nano Banana.
[32:17] content. So, good to see uh Nano Banana. Congrats to Google. Um like great to see
[32:21] Congrats to Google. Um like great to see
[32:21] Congrats to Google. Um like great to see that they are still in the game. I think
[32:23] that they are still in the game. I think
[32:23] that they are still in the game. I think the image stuff, you know, a year ago
[32:26] the image stuff, you know, a year ago
[32:26] the image stuff, you know, a year ago with Sora and um I think it's called
[32:30] with Sora and um I think it's called
[32:30] with Sora and um I think it's called Sora, right? The OpenAI one and and um
[32:34] Sora, right? The OpenAI one and and um
[32:34] Sora, right? The OpenAI one and and um Grock imagine and all that kind of stuff
[32:37] Grock imagine and all that kind of stuff
[32:37] Grock imagine and all that kind of stuff was really fun, really cool. That's not
[32:38] was really fun, really cool. That's not
[32:38] was really fun, really cool. That's not where the value
[32:40] where the value
[32:40] where the value lay for, you know, most people. People
[32:43] lay for, you know, most people. People
[32:43] lay for, you know, most people. People weren't going to pay for that, but a
[32:44] weren't going to pay for that, but a
[32:44] weren't going to pay for that, but a really cool tool. But now we'll see.
[32:46] really cool tool. But now we'll see.
[32:46] really cool tool. But now we'll see. It'll be interesting to see the creative
[32:49] It'll be interesting to see the creative
[32:49] It'll be interesting to see the creative side of things where um people are able
[32:51] side of things where um people are able
[32:51] side of things where um people are able to take this which I think is really
[32:54] to take this which I think is really
[32:54] to take this which I think is really interesting like um just opening up the
[32:57] interesting like um just opening up the
[32:57] interesting like um just opening up the creativity when you remove um the
[33:02] creativity when you remove um the
[33:02] creativity when you remove um the barrier if you will just like with
[33:03] barrier if you will just like with
[33:03] barrier if you will just like with coding. It used to take years to become
[33:05] coding. It used to take years to become
[33:05] coding. It used to take years to become a really good developer. Now anybody can
[33:08] a really good developer. Now anybody can
[33:08] a really good developer. Now anybody can jump on and say hey build me this app
[33:10] jump on and say hey build me this app
[33:10] jump on and say hey build me this app and it builds an app just the same. It
[33:12] and it builds an app just the same. It
[33:12] and it builds an app just the same. It used to take years to become a uh really
[33:16] used to take years to become a uh really
[33:16] used to take years to become a uh really uh you know mature artist uh
[33:20] uh you know mature artist uh
[33:20] uh you know mature artist uh videographer, photographer, musician. Um
[33:24] videographer, photographer, musician. Um
[33:24] videographer, photographer, musician. Um now we remove that barrier to to where
[33:27] now we remove that barrier to to where
[33:27] now we remove that barrier to to where anybody can develop that. But not only
[33:29] anybody can develop that. But not only
[33:29] anybody can develop that. But not only that, this is the the interesting thing
[33:31] that, this is the the interesting thing
[33:31] that, this is the the interesting thing is just like in the developing uh
[33:33] is just like in the developing uh
[33:33] is just like in the developing uh developer use case and coding um uh
[33:36] developer use case and coding um uh
[33:36] developer use case and coding um uh examples with art, there's still a lot
[33:40] examples with art, there's still a lot
[33:40] examples with art, there's still a lot of manual like not fun stuff that
[33:43] of manual like not fun stuff that
[33:43] of manual like not fun stuff that artists have to do to create new art.
[33:48] artists have to do to create new art.
[33:48] artists have to do to create new art. And so you remove that for the artists
[33:50] And so you remove that for the artists
[33:50] And so you remove that for the artists and they become 10 times more prolific
[33:53] and they become 10 times more prolific
[33:53] and they become 10 times more prolific and they know they have taste, right?
[33:56] and they know they have taste, right?
[33:56] and they know they have taste, right? like they know what looks good, they
[33:57] like they know what looks good, they
[33:57] like they know what looks good, they know what sounds good, they know the
[33:59] know what sounds good, they know the
[34:00] know what sounds good, they know the little nuances and details that they
[34:01] little nuances and details that they
[34:01] little nuances and details that they need to fix. Now they can fix those um
[34:03] need to fix. Now they can fix those um
[34:03] need to fix. Now they can fix those um and build build these um really cool art
[34:07] and build build these um really cool art
[34:07] and build build these um really cool art pieces very quickly, which I think is
[34:10] pieces very quickly, which I think is
[34:10] pieces very quickly, which I think is that's the exciting thing to me um is is
[34:12] that's the exciting thing to me um is is
[34:12] that's the exciting thing to me um is is not the slop, of course I I joke about
[34:14] not the slop, of course I I joke about
[34:14] not the slop, of course I I joke about that. Um, it's not just massive amounts
[34:17] that. Um, it's not just massive amounts
[34:17] that. Um, it's not just massive amounts of volume of content, but it's um where
[34:21] of volume of content, but it's um where
[34:21] of volume of content, but it's um where actual artists, just like in the
[34:23] actual artists, just like in the
[34:23] actual artists, just like in the developer world, we see developers
[34:25] developer world, we see developers
[34:25] developer world, we see developers coming up with amazing new ideas that
[34:27] coming up with amazing new ideas that
[34:27] coming up with amazing new ideas that would have never been capable that the
[34:29] would have never been capable that the
[34:29] would have never been capable that the average person is never going to think
[34:30] average person is never going to think
[34:30] average person is never going to think of. Same thing in the art world. I
[34:32] of. Same thing in the art world. I
[34:32] of. Same thing in the art world. I believe that um we're going to continue
[34:34] believe that um we're going to continue
[34:34] believe that um we're going to continue to see artists use these tools to
[34:37] to see artists use these tools to
[34:37] to see artists use these tools to develop um more and more and better art.
[34:39] develop um more and more and better art.
[34:39] develop um more and more and better art. You think about Picasso
[34:41] You think about Picasso
[34:41] You think about Picasso um did created an art piece at least
[34:45] um did created an art piece at least
[34:45] um did created an art piece at least once a day, right? Like he did one art
[34:47] once a day, right? Like he did one art
[34:47] once a day, right? Like he did one art piece a day. Uh extremely prolific and
[34:50] piece a day. Uh extremely prolific and
[34:50] piece a day. Uh extremely prolific and wonderful gift to the world that that we
[34:52] wonderful gift to the world that that we
[34:52] wonderful gift to the world that that we all get to share. And so we'd love to
[34:54] all get to share. And so we'd love to
[34:54] all get to share. And so we'd love to see that in the the Neo artist. That
[34:57] see that in the the Neo artist. That
[34:57] see that in the the Neo artist. That being said, I think traditional
[34:59] being said, I think traditional
[34:59] being said, I think traditional straightup analog human artwork is going
[35:03] straightup analog human artwork is going
[35:03] straightup analog human artwork is going to uh only increase in value as we see
[35:07] to uh only increase in value as we see
[35:07] to uh only increase in value as we see uh these uh kind of AI supported
[35:11] uh these uh kind of AI supported
[35:11] uh these uh kind of AI supported uh um art pieces come out. So, we had a
[35:16] uh um art pieces come out. So, we had a
[35:16] uh um art pieces come out. So, we had a great show today, man. That was our last
[35:18] great show today, man. That was our last
[35:18] great show today, man. That was our last story. Uh I love this. We're having so
[35:21] story. Uh I love this. We're having so
[35:21] story. Uh I love this. We're having so much fun. I hope you guys are getting
[35:22] much fun. I hope you guys are getting
[35:22] much fun. I hope you guys are getting some value out of it. Um I love talking
[35:25] some value out of it. Um I love talking
[35:25] some value out of it. Um I love talking through AI news. what's important, why
[35:27] through AI news. what's important, why
[35:27] through AI news. what's important, why it's important to you. Hopefully, you
[35:29] it's important to you. Hopefully, you
[35:30] it's important to you. Hopefully, you guys are learning something. I'm
[35:31] guys are learning something. I'm
[35:31] guys are learning something. I'm learning something every day, which is
[35:33] learning something every day, which is
[35:33] learning something every day, which is really fun. We're trying to make this
[35:35] really fun. We're trying to make this
[35:35] really fun. We're trying to make this show a little bit better, one day at a
[35:37] show a little bit better, one day at a
[35:37] show a little bit better, one day at a time. If you've got any feedback or
[35:39] time. If you've got any feedback or
[35:39] time. If you've got any feedback or anything that you think we could do
[35:41] anything that you think we could do
[35:41] anything that you think we could do better, uh, send us a note to.live.
[35:45] better, uh, send us a note to.live.
[35:45] better, uh, send us a note to.live. Um, you can also sign up for our
[35:47] Um, you can also sign up for our
[35:47] Um, you can also sign up for our newsletter there, the loop, uh, where we
[35:50] newsletter there, the loop, uh, where we
[35:50] newsletter there, the loop, uh, where we in written form give you some more
[35:52] in written form give you some more
[35:52] in written form give you some more detail. We'll have more kind of
[35:54] detail. We'll have more kind of
[35:54] detail. We'll have more kind of practical application, hands-on stuff
[35:56] practical application, hands-on stuff
[35:56] practical application, hands-on stuff that everybody can do and and use to to
[35:59] that everybody can do and and use to to
[35:59] that everybody can do and and use to to learn this craft and and understand AI
[36:01] learn this craft and and understand AI
[36:01] learn this craft and and understand AI and become wellversed in it. That's our
[36:03] and become wellversed in it. That's our
[36:03] and become wellversed in it. That's our whole goal is that you would become
[36:06] whole goal is that you would become
[36:06] whole goal is that you would become well-versed in these tools and these
[36:07] well-versed in these tools and these
[36:08] well-versed in these tools and these capabilities uh and that it would impact
[36:10] capabilities uh and that it would impact
[36:10] capabilities uh and that it would impact your life positively and for the better.
[36:14] your life positively and for the better.
[36:14] your life positively and for the better. That being said, you come here for AI
[36:17] That being said, you come here for AI
[36:17] That being said, you come here for AI news, real context, and I'm your human
[36:20] news, real context, and I'm your human
[36:20] news, real context, and I'm your human in the loop. Austin Wilson. We'll see
[36:22] in the loop. Austin Wilson. We'll see
[36:22] in the loop. Austin Wilson. We'll see you tomorrow at 6:00 a.m. Central. And
[36:26] you tomorrow at 6:00 a.m. Central. And
[36:26] you tomorrow at 6:00 a.m. Central. And we can't wait for it. Let's go.