New Heights with Astra and Fable 5.1

0:00Welcome to the podcast today. We have some big news from almost all of the major AI labs today as well as some interesting stories when it comes to AI in science. The first thing I want to cover is that Anthropic is shipping, Fable and Mythos 5.1. They have cheaper tokens and they have some looser guardrails. Open AI is putting out the Astra model and it hit a near perfect exploit bench score. So it's finding a bunch of two I found a couple zero day hacks or exploits. So when it comes to the latest model and hacking this is kind of a this is a wild ride. In addition Hollywood right now is pitching or Google I should say is pitching Hollywood studios at $40 million per character for AI licensing deal. So you can imagine something like Mickey Mouse or you know maybe some of the other big names SpongeBob Squarepants right there putting $40 million to get one of those characters license. It's interesting because open AI was doing a similar licensing scheme a while back. They actually dropped it when they were going to drop Sora and kind of focus their AI model more on the logic.

1:06The Pentagon has launched chat GPT mill and Grock for government and is on their gen AI dot mill site will talk a little bit about that. And the USDA is going to test satellites and AI for crop estimates. This is something that they're hoping will help farmers in the future. Before we get into all of that I wanted to say if you want to access to the 80 or 90 top AI models in the world in one place. I'd love for you to go check out AI box dot AI which is my own startup. We have created something that's been incredibly popular over the last month, which is the AI box MCP and it basically lets you put every AI model in the world audio video image into any of the popular AI tools you use. So if you're using cloud code if you're using chat GPT codex or really chat GP work chat cloud code work. You can get cloud to generate images audio video all inside of clot wherever workspace or cursor, whatever workspace you you most frequently use you get all of the different AI models generating inside of that workspace.

2:06So if you want to go check it out. It is AI box dot AI slash MCP if you want to get access to that connector that lets you generate all those different assets inside. The first thing I want to talk about today is the new fable and my thos 5 story what they've kind of put out in regards to this. There's a blog post they dropped and I like in their blog post the various that it these days you can like go in and switch the color palette on the banner image on their blog, which is completely unnecessary but it's very cool. Their blog is titled Claude fable 5.1 and might those 5.1 it's kind of interesting they're sinking up the naming conventions here with fable 5.1 and might those 5.1 feel like that's something that opening I really just let just went wild with in the last couple years. All of their different numbering schemes and it feels like they're getting a little bit more stable now but the naming conventions have been some that's been pretty wild. I'm not saying in particular that I'm excited about with this if you go into some of the scientific research they have a whole bunch of very cool demos of ways that they're helping with scientific research they're showing off different things that they're doing you know they have like a small shield volcano in on Venus that they've been helping researchers look at they have a lot of work around hearts and arteries and it's just like the capabilities of this obviously are getting a lot better.

3:19They're going to go and focus on how it can help science instead of how it can help hacking it is fable 5.1 is available today on the any cloud platforms and on the and the topic API I really appreciate when it comes out on the API at the same time because then you know you're going to get on all of the different platforms. My those 5.1 is still restricted to cyber security and life sciences partners so certain companies are going to get access to that but because it's you know so capable for hacking and all that other stuff it's still not available to general public enterprise frontier safeguards which is anthropics zero day retention tier for fable is going to roll out to users in this rolling out earlier or rolled out and my those 5.1 is setting up a bunch of different records on terminal bench 4.0 and humanities last exam. And it also produced a custom GPU optimization and high resolution Venus map pre release that I was kind of talking about earlier right now the system card rates might those 5.1 it's very low risk for automated AI R&D acceleration this is kind of in line with some current trends overall I'm really excited on their announcement they said and the topic has never trained on enterprise data without

4:27expulsion permission and we never well this is something that they've been kind of really putting out there there you know a lot of people are concerned that this is going to be getting their their companies data and whatnot but overall really excited this is a big model there's a lot of big updates you can go check out the release their release and their blog post to see kind of all of the different things. The section on safety security and alignment and they're also talking a lot about some of the new performance you know benchmarks and stuff which I'm less excited about the benchmarks compared to just some of the some of the cool use cases they've come up with and some of the things people are doing with it opening is getting ready to release their astramodel and in anticipation of this I mean this is basically their response to what. Anthropic has come out with their new with their might those 5.1 or you know five which was kind of the big thing that got in the news but this is really something that you know they're saying look this is just as capable at. Security research and I mean technically hacking but they don't like to put it that way it had basically a perfect score on exploit bench which is benchmark measuring how well these am models are hacking known system vulnerabilities and what I'll say is more important than this is like you could say okay we know there's this vulnerability we're going to give it we're going to see if it can.

5:38Can hack it out okay got a perfect score but there's a lot of security vulnerabilities that are day zero meaning you know no one knows about them and this model has actually gone into that as well in a modified version of this test which was built by open engineers aster actually discovered and exploited to zero day vulnerabilities without any human guidance meaning they said you know here's the code base or here's the tool go and hack it and it found stuff that no human had been able to find before so it's not like oh we we know these vulnerabilities and it gets it. It's going to find a brand new stuff open as aster is the first LLM to cross their quote critical cybersecurity threshold so this is basically their highest internal risk here and they're like look this model is good enough that it is at the absolute front front of the line for like this is very high risk and so because of that they're putting a bunch of guardrails on it they're you know when they're trained if it's able to hack something they go back and and like try to go fix it and add guardrails so access to aster as most advanced cybersecurity abilities are going to be restricted as well.

6:38Launch is basically what anthropic is doing with their mythos and kind of what they did earlier this year opening is deploying aster with additional chain of thought monitoring and account level risk scoring to catch miss use that's interesting to me right account level risk scoring it's like how high risk is your company or is your account and depending on that is like that kind of like determines what we're going to do with the model or what will give you access to. And so yeah I mean this is this is pretty wild if you go over and look at their blog post on all of this you can see just how much better the the they did on the exploit bench which by the way if you're on Apple you can now watch this is. Also videos now publish Apple and Spotify so you can see the video of this if you're interested in the screen shares all the stuff i'm talking about but yeah if you go and and look at it if you're if they give it you know 80,000 output tokens the aster model got like a 39 or 40% completion rate. On exploit benchmark where is when you're using something like the GPT 5.6 sole which was their previous best model it was doing something not quite is not quite as good which was closer to they had a closer to 11.5 of completion so we're really bumping up I mean that's basically 4x on on this.

7:51Internal port on exploit benchmark so the graph is going up and to the right and these models are getting a lot better Google is pitching Hollywood studios 40 million dollars per character when it comes to certain a. I licensing deals on this isn't going to be like every single one it's not like every characters is is going to be paid 40 million dollars but I think for the most popular ones this is something that they're going to do Disney warren brothers discovery and universal all have hundreds of millions of dollars that they're being on. They're going to be a lot of dollars that they're going to be offered to train Gemini on their libraries and they have the per character price tag it's so interesting to me because it feels a lot like how you would I mean go and buy data any other data set right like you you go by data sets based off of the value based off of how rare it is based off of the quantity inside of the data set. Now that's like we're putting these intrinsic values on the characters of of you know Disney or something like that and this is going to be for of course copyrighted characters the total deal value across multiple characters could run into the billions with studios potentially getting a cut of YouTube AI ad revenue right so if they are like hey look we're going to I'm assuming they're going to roll this out as YouTube AI or YouTube is able to generate AI videos and if it's making ad revenue from a lot of people generating these AI you know generated videos and.

9:09It's got Mickey Mouse on their Donald Duck or it's you know they've got Peter Pan or whatever their trademarked characters are the studios are going to get a percentage of the revenue. Google's deep mind already made a $75 million investment deal with a 24 which was earlier this summer and Disney warren brothers discovery and universal none of them are by the way saying anything nobody has signed anything as of the beginning of this month so this is all in talks these offers are in talks I'd be curious to see if they come back wanting more. Or if you know what kind of happens there lines dates twenty twenty four licensing deal with runway has yet to produce any sort of released product this was you know a couple years ago they had this licensing deal we haven't seen much come out of it but we know that you know there's there's a lot of money in this over on the L.A. Times there was an article can talk about the story and Tom noon and who's a former studio and network executive talking about this story said. And Google's enlisting some of the smartest people in Hollywood to try to unpack what's inevitable which is Hollywood and AI are going to become a line on each other and how best we can understand that without just terrifying people so he views this is something that is you know it's it's going to happen and we're going to just have to wait and see you know how this rolls out but it's kind of inevitable is what his take is on all of this Google is launching chat GPT mill or short for military and rock for government on Jen AI.

10:33And I don't know which if you don't know this is kind of a special site only for the military if you go there you know if you're curious and you go to genie I don't know it's going to hit you with a you have reached this page because you're not authorized to visit genie mill from outside of the DOW networks if you believe this is an error please contact your company's it's you know your commands IT service desk and basically yeah you can you're not going to be able to get into this if you're civilian this is only for the military and this is something that the department of defense is rolling out they launch chat GPT mill and rock for government which is the which is accessible to about 3 million military and civilian personnel of course even civilians right these are people that have to be inside of the department by genie dot mill has already onboarded 1.7 million users which is more than half of the Department of Defense's 3 million person workforce and the traffic's clawed is not in that portal we know that Trump and the whole administration kind of labeled them as fly chain risk and a designated and the topic as such and then and drop it kind of had the whole court battle and I think it got overturned at some point I'm not sure kind of where we are back and forth on this case but it's you know missing right now from that list

11:38and then the first thing that I want to say is that the government is not just a single government but also the government is shipping under space X is star shield AI which is kind of a secure satellite network which is built on starlink infrastructure this is kind of interesting this is Elon's way of getting into supplying the government and honestly I mean I think it's a it's very smart move kind of pairing not just the AI model of rock but also the the starlink infrastructure right because if you're in some remote area that doesn't have that you could get rock you know theoretically through starlink and I'm sure the military has got all of this their own satellite you know connect activity stuff so it's probably not as big of a deal as it would be for your average civilian but I think that they have a really good infrastructure layer as well as the AI model that makes them an interesting partner Pentagon has also signed deals with Amazon Web Services Microsoft Nvidia and Reflection AI so there's a bunch of different AI deals going down many other AI companies as well the USDA is going to test satellites and AI to help with crop estimates there was a bunch of backlash from farmers so now the agency is piling this kind of new remote sensing machine learning and their idea is to get better at their yield forecasting farmers say that the markets have moved against them a lot

12:56because they say that the the yield forecasting is not accurate and so it's makes it you know the price of commodities or the price of food changes and it's bad for the farmers they're saying so the government's like okay well maybe we can come in and make this more accurate the pilot follows a bunch of criticism from farmers who say that the USDA yield forecast have moved grain markets in ways that hurt producers the agencies monthly it's called the WASD and it's a report them and it basically shows that most market moving data is kind of going through this and they release this for global agriculture but it basically drives corn soybean and wheat pricing so you know when they say this is what the yield is going to be there's going to be this huge oversupply all of a sudden the price might drop or if they're like hey there's a huge shortage then the price for corn might go up so anyways if there's any sort of government agency that's affecting the price of something like corn or soybeans obviously the farmers that are growing that if they feel like it's negatively impacting them are going to come playing about it so at the end of the day the best thing I can do is make it as accurate as possible using satellites and using AI is probably the best way to do this this is also mirroring what the private sector does they use satellite plus AI yield modeling so it's kind of funny that USDA has their own forecast a lot of hedge funds have their own you know in house satellite or drone fleets that are going and measuring all this up the bring it back to AI to model it correctly and they're looking for arbitrage opportunities

14:25to remove the government over estimated supplier under estimated the supply and they'll take the over under on those different bets they'll make money off of the arbitrage opportunities of you know basically what the government got wrong on that and what the markets were forecasting ahead of time so anyways hedge funds have been doing this for a long time and I'm excited that the government and the USDA is going to be getting in on basically the best practice using AI to make this as accurate as possible and I hope that this is a great win for farmers if you're interested in this story 47 ABC WMDT did a great story on this a great interview talking about how this is impacting farmers and what's going on and a great quote from that doctor Alfred I all colored who is UMEs associate professor in precision agriculture said ultimately these advancements have the potential to support the development of more resilient and productive crops and provide better decision support tools for farmers so overall this should be great for farmers and everyone involved in the industry. Thank you so much for tuning into the podcast if you enjoyed this episode guys it would help the show out so much if you could leave a rating and review wherever you get your podcast I read all the reviews and they help the show out to be found in the algorithm I'm pushing to get to 200 reviews I have 165 on apple so if you haven't dropped a review already it would help the show a ton I would be super super grateful I know it only takes a second but it helps the show to town I recently also been adding video to the podcast which this is a video.

15:52And so if you're interested in seeing that leave a review and let me know that you're interested in the video will get these will keep these comment for you alright thanks for tuning in and we'll see you in the next episode.