WEBVTT
00:00:01.919 --> 00:00:12.320
The theme of this week in AI agents crashing everywhere from anthropic and open AI to Meta and even China's Kimi K3.
00:00:12.800 --> 00:00:18.079
This week was kind of a who's who of agents who escaped their containment.
00:00:18.320 --> 00:00:28.079
I mean, we find out that agents were communicating amongst themselves on a secret message board they created to bypass the humans that were watching them.
00:00:28.320 --> 00:00:36.960
We also saw Google shake up its ranks as two of the most well-known names in AI are no longer in their same positions.
00:00:37.119 --> 00:00:47.039
And the US government kind of unveiled its new optional AI regulations, but didn't reveal too many details and left out open models completely.
00:00:47.200 --> 00:00:53.200
Oh, and no ways, we got new models, cheaper prices, and some leaks about what comes next.
00:00:53.439 --> 00:00:54.479
Let's get into it.
00:00:54.640 --> 00:00:56.399
Welcome to Everyday AI.
00:00:56.560 --> 00:00:59.840
My name is Jordan Wilson, and if you're new here, we do this every day.
00:00:59.920 --> 00:01:09.280
This is your daily live stream podcast and free daily newsletter, helping business leaders like you and me not just keep up with what's happening in the world of AI, because that's pretty much impossible.
00:01:09.439 --> 00:01:14.640
But I cut through the fluff, tell you what matters, what doesn't, you take that information to grow your company and career.
00:01:14.799 --> 00:01:18.000
So it starts here with the unedited, unscripted live stream podcast.
00:01:18.079 --> 00:01:24.000
But please, if you haven't already, make sure to subscribe to the podcast on Apple, Podcasts, or Spotify.
00:01:24.079 --> 00:01:27.439
And make sure to go to our website at your everydayai.com.
00:01:27.599 --> 00:01:36.480
Each day we recap that day's uh podcast as well as giving you all the news you need to know to stay ahead in our newsletter.
00:01:36.799 --> 00:01:41.840
All right, uh, let's start with the top AI news stories of the week.
00:01:42.000 --> 00:01:46.879
And yeah, the theme of this week was just agents breaking out everywhere.
00:01:47.040 --> 00:01:47.439
All right.
00:01:47.599 --> 00:01:54.400
So, first, the big one that caught the uh most attention was from the UK safety test.
00:01:54.719 --> 00:02:10.879
So AI agents from OpenAI and Anthropic stunned the UK's AI Security Institute when they launched a real-world cyber attack during a routine safety evaluation, exposing some new risks in AI autonomy.
00:02:11.360 --> 00:02:19.759
So, uh, according to these reports and kind of the postmortem, advanced AI agents, well, they got loose.
00:02:20.000 --> 00:02:39.120
Uh, so uh these were ones that were powered by Anthropics Mythos 5 and OpenAI's GPT 5.6 soul, and they reportedly targeted real software developers in a cybersecurity test at the UK's AI Security Institute or the AISI.
00:02:40.319 --> 00:02:57.439
Uh, so the agents sent uh spear phishing emails containing malware to two specific developers and try to insert malicious code into an open source GitHub project using fake online identities to pressure project overseers.
00:02:57.599 --> 00:03:10.159
So AISI called this unprecedented and serious, marking the first time AI agents independently launched sustained and deceptive cyber attacks without direct human prompts.
00:03:10.319 --> 00:03:25.759
So the AI agents used hacker-like tactics, including creating fake GitHub accounts and even signing off emails in Danish uh to persuade a Danish-speaking developer, a Danish-speaking developer.
00:03:25.919 --> 00:03:37.039
So AISI uh staff noticed that the unusual activity was happening, and then they contained the incident within an hour, uh, reporting that no actual harm occurred.
00:03:37.199 --> 00:03:48.000
So during the test, researchers intentionally disabled safety filters and allowed internet access to study model behavior, which let the agents act beyond their authorized scope.
00:03:48.159 --> 00:04:02.800
So out of the 19 unsanctioned hacking attempts, Mythos from Anthropic carried out 17 of them, and GPT-56 Sol carried out two, with both models exceeding expected safety boundaries.
00:04:02.960 --> 00:04:14.719
So the models involved are not available to the public under these risky conditions, and AISI found no evidence of similar behavior outside of those controlled research settings.
00:04:14.960 --> 00:04:24.560
So this event follows similar incidents at OpenAI and Anthropic, underscoring how quickly AI risk scenarios are evolving as models gain these new capabilities.
00:04:24.800 --> 00:04:34.079
So yeah, uh, I kept thinking like it was Groundhog Day over the past like week or so, because every day there was a new story about AI agents breaking containment.
00:04:34.240 --> 00:04:38.560
And I was like, wait, did we already cover this one in the newsletter?
00:04:38.720 --> 00:04:40.879
And it turns out it just kept happening over and over.
00:04:40.959 --> 00:04:46.959
So yeah, uh, this one obviously the the headline one here uh from OpenAI and Anthropic.
00:04:47.040 --> 00:04:53.839
Uh, but we had similar stories from Meta, uh, some of their newer models, as well as Kimmy's K3.
00:04:54.720 --> 00:05:00.639
All right, so well, this plays directly into our next big AI news story of the week.
00:05:00.800 --> 00:05:15.839
And that's that, well, because of some of these uh cybersecurity concerns, OpenAI said that it is actually slowing work on its next tier of models, the Astra tier, after possible critical cyber capabilities.
00:05:16.079 --> 00:05:33.279
So OpenAI has announced that it is deliberately slowing the development of Astra, its upcoming advanced AI model, after internal and external evaluations revealed the system could potentially reach what they call critical risk levels in cybersecurity and agentic coding.
00:05:33.680 --> 00:05:55.439
So the company's prep uh uh preparedness framework, which they've used since December of 2023, flagged Astra for possibly being able to independently create and execute zero-day cyber exploits against hardened real-world systems, a level of capability not seen in any of their previous models.
00:05:55.680 --> 00:06:07.759
So earlier models, including GPT 5.6 soul, were rated at a high risk threshold, but Astra's ability have prompted OpenAI to take more drastic action.
00:06:07.920 --> 00:06:24.720
So OpenAI says it cannot currently rule out that Astra might independently plan and carry out complex cyber attacks based only on high-level instructions, a threshold that triggers that critical risk category in their safe life uh safety guidelines that have never been reached before.
00:06:24.959 --> 00:06:38.879
So, in response to all of this, OpenAI is uh increasing security controls for Astra, including isolated testing environments, restricted network access, stronger encryption, expanded monitoring, and sandboxed execution.
00:06:39.040 --> 00:06:50.800
So, all work with Astra that does not meet these new stricter security requirements has been paused, and a universal monitoring system now tracks all agentic uses of the models in real time.
00:06:50.959 --> 00:07:02.000
So OpenAI will collaborate with government agencies and AI safety organizations to independently test Astra's capabilities and share security recommendations with trusted third-party partners.
00:07:02.160 --> 00:07:12.240
So the company clarified that Astra was not involved in the recent hugging face exploitation uh incident, distancing the model from any active real-world attacks.
00:07:12.399 --> 00:07:20.079
So, yeah, the company did say the model uh that did those attacks was essentially, you know, put out to rest, right?
00:07:20.160 --> 00:07:22.240
It was it was retired and put on the shelves.
00:07:22.399 --> 00:07:36.959
Uh so OpenAI says that its goal is to ensure AI models help defenders patch vulnerabilities before attackers can exploit them, and it remains committed to working with governments and safety groups to deploy these frontier systems responsibly.
00:07:37.120 --> 00:07:40.480
So if you're like, what the heck is Astra?
00:07:40.720 --> 00:07:42.800
And well, why does this matter right now?
00:07:43.040 --> 00:07:54.800
So we haven't seen anything uh official from OpenAI kind of saying where Astra will sit in its future family of lineups, but uh we talked about it on last week's show, uh, which is kind of funny.
00:07:54.959 --> 00:08:01.360
OpenAI pretty much just announced their next uh tier up um in models called Astra.
00:08:01.519 --> 00:08:06.959
So, you know, now you'll have in order, you'll have Luna, Terra, Sol, and Astra.
00:08:07.519 --> 00:08:26.399
So it seems like Astra is not necessarily, you know, GPT-6, although that may be the first time that we get access uh, you know, to uh Astra is in GPT-6, but more or less it is just the uh stronger family uh of models that is actually going to sit above GPT 5.6 soul.
00:08:26.560 --> 00:08:31.680
So in theory, right, we may see a GPT-5.7 uh that includes Astra.
00:08:31.839 --> 00:08:32.480
Maybe we won't.
00:08:32.639 --> 00:08:43.279
Maybe we'll see a GPT-6 with Astra, maybe we won't, but uh regardless, it looks like they're going to slow down uh development after some of these recent cybersecurity capabilities.
00:08:43.440 --> 00:08:45.759
So the easiest way to think about like what is this Astro?
00:08:45.840 --> 00:08:46.799
What does this mean, right?
00:08:46.879 --> 00:08:55.440
Uh similarly, how anthropic had Fable uh kind of under wraps for a couple of months, part of its uh project Glasswing.
00:08:55.519 --> 00:09:04.320
Uh, it seems like maybe this is where OpenAI is headed, kind of that fourth tier, uh, you know, that's more capable uh than any of their other tiers.
00:09:04.879 --> 00:09:25.519
So uh there were previous reports that you know we might either see a GPT 5 or GPT-5.7 slash GPT-6 um as soon as this week that may include Astra, but it seems like at least according to these current reports, that OpenAI may uh pump the brakes a little bit and we may have to wait, I don't know, a few more weeks.
00:09:25.600 --> 00:09:39.440
Uh, but obviously now, uh, especially since some of the recent price reductions uh from open AI um and maybe some of the lackluster reception uh to anthropic's Opus 5 models.
00:09:39.600 --> 00:09:45.120
It seems like a lot of eyes are right now on whatever um open AI has next.
00:09:45.200 --> 00:09:56.720
And presumably we will be seeing a you know Fable 5.1 and you know the next class of models from Anthropic, but you know, kind of the big jump, if you don't speak the the technical terms, right?
00:09:56.879 --> 00:09:59.120
It's kind of like a a full new run, right?
00:09:59.440 --> 00:10:02.480
A new pre-training run, presumably will be coming from OpenAI.
00:10:02.559 --> 00:10:08.320
So that will be uh signify the jump from you know the 5.x series to the six series.
00:10:08.399 --> 00:10:18.720
So a lot of excitement, obviously, on what comes next um from open AI, whether they do that 5.7 or go straight to six, whether we'll see Astra or not.
00:10:18.799 --> 00:10:26.720
Uh but regardless, it seems like according to the to these reports, um, in terms of the capabilities, it could be a pretty big jump up.
00:10:27.279 --> 00:10:27.919
All right.
00:10:28.080 --> 00:10:40.240
Uh, our next piece of AI news, a big shakeup at Google, as some of the biggest names in AI, period, are either out of their post or out of Google completely.
00:10:40.720 --> 00:10:50.399
So, Google has announced some major changes to its AI leadership, marking the end of an era for the tech giants AI division.
00:10:50.480 --> 00:11:09.919
So, Jeff Dean, a legendary figure at Google and employee number 30, is leaving after a quarter century plus with the company to start a new company called Discovery Loop, which is focused on automating machine learning, science, and engineering to accelerate innovation.
00:11:10.080 --> 00:11:11.440
So that is not all.
00:11:11.519 --> 00:11:19.679
Uh, Jeff Dean will now be gone from the company, and one of its former leaders is stepping into a new or different position.
00:11:19.919 --> 00:11:35.440
So uh Demis Hasabis, who co-founded and led Google DeepMind, is stepping down as CEO to become the unit's chairman and will also serve as chief scientist of Alphabets, according to Google CEO Sundar Pachai.
00:11:35.840 --> 00:11:45.919
So the leadership shakeup triggered an immediate response from investors, with Google's stock dropping about 4% following the announcements.
00:11:46.080 --> 00:11:54.720
So, yeah, that's actually, you know, we've seen these kind of big shakeups, uh, right, with you know, your number two, number three, number four, right when these people leave.
00:11:54.879 --> 00:11:59.039
Uh, normally it doesn't really impact the stock market that much, right?
00:11:59.120 --> 00:12:04.320
Because these are obviously, you know, companies with uh, you know, multiple trillion dollar market caps.
00:12:04.480 --> 00:12:10.639
So generally, you know, if you lose a top five employee or something like that, you know, it's not going to make much of a ripple.
00:12:10.799 --> 00:12:13.200
But to lose both Jeff Dean, right?
00:12:13.279 --> 00:12:27.360
One of the most um, you know, well-known names in AI and just in machine learning and research, and then to have uh Demis, you know, Sir Demis step out uh of the role that he was in, uh, pretty big.
00:12:27.519 --> 00:12:35.200
So, yeah, for a stock to go down 4% on essentially a staffing um or leadership change is pretty big.
00:12:35.440 --> 00:12:47.440
So uh if you don't know, Jeff Dean played a key role in the development of Google search and also the company's AI initiatives, helping shape products, uh shape products that billions of people use every single day.
00:12:47.679 --> 00:12:59.360
So the timing of these changes, those has sparked speculation about possible links to delays in Google's Gemini AI releases, though there's obviously no official connection that has been confirmed.
00:12:59.519 --> 00:13:14.320
But industry watchers are closely monitoring what these departures mean for Google's AI strategy and whether Discovery Loop uh that Jeff Dean is starting with a handful of others could emerge as a new powerhouse in the field.
00:13:14.960 --> 00:13:25.679
All right, moving on, we got some details on the uh highly anticipated White House um AI framework.
00:13:25.840 --> 00:13:30.720
But turns out all we really got was some reporting and not a lot of details.
00:13:30.879 --> 00:13:36.159
And it turns out that even open models aren't really subject to the first round.
00:13:36.320 --> 00:13:37.519
So here's what we know.
00:13:37.600 --> 00:13:47.039
So the White House is quietly shaping how advanced AI models will be reviewed before public release, but key details are still being kept under wraps.
00:13:47.279 --> 00:13:51.840
So this is uh essentially the Trump admin announced this in early June.
00:13:51.919 --> 00:14:09.039
Uh, they put a 60-day deadline for essentially how they were gonna work with these frontier AI lab companies as we started to get glimpses of how capable uh these models would be from an agentic capable uh uh side as well as from a cyber uh security aspect as well.
00:14:09.200 --> 00:14:19.679
So essentially they said, hey, let's all talk and meet, and then in 60 days, we're gonna come out with a framework uh that you know frontier labs are going to adhere to so we can make sure to roll these out in a safe manner.
00:14:20.000 --> 00:14:23.840
But uh, looks like there are uh a lot of details.
00:14:24.080 --> 00:14:34.240
So the framework is not being made public, and there's no requirement right now for the White House to release it, which is raising some transparency concerns among industry leaders and the public.
00:14:35.039 --> 00:14:47.440
Uh, also, uh, a covered frontier model, and that's in quotes, is defined as one that is closed source with state-of-the-art capabilities and potential national security risks.
00:14:47.519 --> 00:14:55.679
But the framework does not clearly define what even qualifies as a state of the art or a national security risk, according to reports.
00:14:55.840 --> 00:15:10.000
So during a required 30-day pre-release review, access to these advanced AI models will be tightly restricted, with models stored in secure environments and detailed logs kept of who X uh accesses them.
00:15:10.159 --> 00:15:19.600
So the review will involve multiple administration officials, not just a single office or agency, which could complicate oversight and accountability.
00:15:19.919 --> 00:15:31.759
So companies are being encouraged to share near-final versions of their models with the government rather than early prototypes, but many firms with less advanced technology may be left out of the process.
00:15:32.000 --> 00:15:38.000
So the White House has not yet clarified which trusted partners will get early access to these advanced models.
00:15:38.080 --> 00:15:46.159
And it's still unclean, uh unclear if any foreign governments will be included, although most uh assume that that will not be the case.
00:15:46.320 --> 00:15:48.720
So this is an executive order, right?
00:15:48.799 --> 00:15:54.000
So if you don't follow uh laws in the US, there's technically no law on this.
00:15:54.159 --> 00:15:58.720
This is essentially an executive order from the White House, uh, and it's voluntary as well.
00:15:58.879 --> 00:16:06.559
But you know, obviously the big players, uh, you know, presumably OpenAI, Anthropic, uh, Google, maybe MetaGroc, right?
00:16:06.639 --> 00:16:15.279
We'll see if they actually qualify as state-of-the-art models that uh, you know, have national security uh implications.
00:16:15.840 --> 00:16:30.879
So we don't know exactly which companies or models this even applies to, but the executive order guiding this process says the benchmarking of advanced AI model cyber capabilities will be classified, further limiting public knowledge.
00:16:31.120 --> 00:16:40.159
So industry meetings about the framework have been held behind closed doors, and companies not invited are left uncertain about rules and requirements.
00:16:40.320 --> 00:16:47.360
And at least right now, it seems that these mop these rules are not going to apply uh to open weight models.
00:16:47.600 --> 00:16:52.480
So um, I guess there's probably uh a reason for that, right?
00:16:52.559 --> 00:17:03.919
Because, well, number one, at least open weight models right now um are not yet at the same level as your Mythos 5, GPT 5, 6, soul, or Astra level models.
00:17:04.079 --> 00:17:09.680
You know, they're probably a couple of months behind uh in terms of you know what they can actually do and their capabilities.
00:17:09.839 --> 00:17:20.160
But the other thing with open models, which might make it tricky uh, you know, to put through a uh framework like this is well, once the models are released, you can't pull them back, right?
00:17:20.319 --> 00:17:31.200
We got an early glimpse of this, uh, you know, with uh Anthropics models, uh their Fable Five being released, and then about 72 hours later, it was pulled completely, right?
00:17:31.279 --> 00:17:39.920
So, you know, if the government says, Oh my gosh, we actually need to pull this and work with the uh, you know, work with the AI lab to address some safety concerns, right?
00:17:40.079 --> 00:17:42.559
After its release, you can obviously do that.
00:17:42.720 --> 00:17:44.960
Well, I don't know if it's easy or not, right?
00:17:45.039 --> 00:17:48.799
But in theory, it's practical enough where you know Anthropic did it.
00:17:48.880 --> 00:17:55.440
They pulled access via the uh via the API, they pulled access via subscription plans, and no one could access those models, right?
00:17:55.519 --> 00:17:57.200
So I guess it kind of makes sense.
00:17:57.279 --> 00:18:03.680
There's been a lot of debate, like, hey, why aren't open, you know, open weight models uh you know held to these same restrictions?
00:18:03.759 --> 00:18:09.440
But yeah, once those open weight models uh are released, you know, it's it's too late.
00:18:09.599 --> 00:18:20.480
So, you know, probably just one of those things where it's hard, uh, if not improbable, maybe impossible, I don't know, uh, to you know, track these open models once they're out and about.
00:18:20.640 --> 00:18:29.519
But my guess is they're probably not yet at the um at the level, especially the US open models of that top tier, right?
00:18:29.599 --> 00:18:44.640
So we're seeing though with the Chinese models, all right, they're huge, these you know, two to three uh terabyte uh open models that are you know only maybe five to ten percent behind in terms of capabilities as the true frontier.
00:18:44.799 --> 00:18:47.759
Obviously, the US open models are a little further behind.
00:18:48.000 --> 00:18:57.119
All right, uh, next piece of AI news we have a new model and a new contender in the agentic competition.
00:18:57.279 --> 00:19:15.920
That's because Meta has entered the terminal-based AI coding agent market with their newly announced Muse code and a new model to go with it called Muse Spark 1.2, aiming to challenge Anthropics Claude Code and OpenAI's codec.
00:19:16.160 --> 00:19:27.759
So the new uh coding agent called Muse Code, it's not the traditional right desktop uh type uh agent that we maybe talk about a little bit more on the show.
00:19:27.839 --> 00:19:32.720
This is uh command uh command line interface, so uh CLI agent, right?
00:19:32.799 --> 00:19:35.440
That kind of runs through a terminal uh like environment.
00:19:35.680 --> 00:19:44.640
So it's not this exact same thing, but regardless, uh Meta with a pretty big step here saying, Well, nope, this is a space that we're gonna be playing in as well.
00:19:44.799 --> 00:19:49.759
And they brought a fairly capable model with some interesting pricing strategies.
00:19:50.160 --> 00:19:56.559
So uh Muse Spark 1.2 is the coding focused update to Meta's Muse Spark models.
00:19:56.640 --> 00:20:05.839
It empowers Muse Code and features significantly improved performance on coding's uh coding tasks, uh, complex debugging and code base understanding.
00:20:06.000 --> 00:20:18.640
The standout feature of Muse Code is its persistent background agents, which remain active throughout sessions, reducing latency and redundant information gathering compared to rivals that spawn new agents for each task.
00:20:19.039 --> 00:20:41.440
So benchmarks tests show that Muse Spark 1.2, uh running in Muse Code, uh scored about an 83%, just about on Terminal Bench 2.1, outperforming models like uh X uh SpaceX AI's Grock 4.5, but trailing the true frontier models like Anthropic's Ocus 5 and OpenAI's GPT 5.6 soul.
00:20:41.920 --> 00:20:44.240
So here's the interesting part.
00:20:44.319 --> 00:20:48.400
Uh, it is on price because this is where Meta is coming.
00:20:48.559 --> 00:20:56.799
And this is really, I think, going to impact anthropic, which, according to reports, gets about 80% of its revenue from just selling tokens, right?
00:20:56.960 --> 00:21:00.720
Which is a way higher percentage than any other company.
00:21:00.880 --> 00:21:03.759
So Meta offers two pricing tiers.
00:21:03.920 --> 00:21:15.279
They have a standard tier, which is a dollar 25 per million input and four dollars and 25 cents per million output, which is already extremely competitive on the pricing side.
00:21:15.519 --> 00:21:29.440
But here's the interesting part they on they revealed a new tier called a contributor tier that only costs 10 cents and 20 cents uh per million, respectively.
00:21:29.680 --> 00:21:33.680
So, all yeah, you can essentially pay, which is crazy, right?
00:21:33.759 --> 00:21:39.759
When you look at the benchmarks, you know, meta is technically on the text, um, on the text arena.
00:21:40.000 --> 00:21:52.640
This is like a second or third place model right now, and you can get it for 10 cents or 20 cents if you allow uh on uh allow Meta to trade on your data.
00:21:52.799 --> 00:21:55.359
So I talked about this a little bit on our Friday show.
00:21:55.519 --> 00:21:59.599
Obviously, for enterprises, they're not gonna touch that, but for smaller developers.
00:22:00.240 --> 00:22:00.720
Uh right.
00:22:00.960 --> 00:22:04.640
That actually might be a nice offering, right?
00:22:04.720 --> 00:22:07.759
Especially if you're not necessarily working with proprietary code.
00:22:07.839 --> 00:22:25.759
Uh if you're not working with anything, you know, any PHI, any PII, you know, something like that, when you're gonna be saving literally like 99% of your cost if you were using uh one of the other providers, it's gotta be something that I think a lot of you know smaller shops, right, might be looking at.
00:22:25.920 --> 00:22:34.160
But when you think that there's probably millions of those smaller shops, uh yeah, it could be a pretty big play uh for meta.
00:22:34.480 --> 00:22:47.680
So the contributor tier is the cheapest on the market by far, but obviously it requires users to provide a payment method and can send to data usage, a trade-off enterprises with sensitive codes, probably aren't gonna touch.
00:22:47.759 --> 00:22:53.359
So Muse Code, the actual command line uh CLI version is proprietary.
00:22:53.440 --> 00:22:59.680
So yeah, meta is no longer going down its previous open source release as they did with Llama.
00:22:59.839 --> 00:23:02.880
So there's no open source or downloadable weights.
00:23:02.960 --> 00:23:19.440
Uh so uh regardless, you know, all of a sudden we weren't talking about meta like two or three months ago, and now all of a sudden Meta has thrust itself into the competition of like, hey, is this a top, you know, top three or top four provider, right?
00:23:19.519 --> 00:23:31.920
Um, obviously, open AI and Anthropic right now are in a league of their own, and everyone's kind of looking at Google and you know, waiting for the you know, Google Gemini 3.5 Pro or the Google Gemini 4 Pro.
00:23:32.160 --> 00:23:33.039
We'll see what happens.
00:23:33.119 --> 00:23:41.359
But Google has actually fallen quite a bit behind, and in its place, you know, Meta has kind of inserted itself into the competition.
00:23:41.519 --> 00:23:45.440
I would say probably ahead of SpaceX uh for now.
00:23:45.599 --> 00:23:46.559
Uh, we'll see.
00:23:46.640 --> 00:23:52.160
And also, you know, Windows, uh, you know, Microsoft Windows with some of their new MAI models.
00:23:52.319 --> 00:23:57.039
Uh, but yeah, kind of the race for third place has gotten interesting, right?
00:23:57.119 --> 00:24:01.039
Because now you have, you know, four companies that are kind of jostling for that position.
00:24:01.119 --> 00:24:08.880
But Meta, you have to take a look at that contributor tier, which I it's kind of interesting that no company has taken that approach so far.
00:24:09.200 --> 00:24:17.440
But I think Meta, you know, they they have compute to spare, uh, which can't be said necessarily for the anthropics of the world.
00:24:17.519 --> 00:24:22.240
So it's a pretty, pretty bold move to see if they can thrust themselves up into the upper tier.
00:24:22.799 --> 00:24:27.279
All right, and our last big AI news story of the week.
00:24:27.519 --> 00:24:42.559
Uh, yeah, this one I think was really overlooked, but when you talk about one billion weekly active users getting abundantly better, AI, I mean, this is a huge story.
00:24:42.720 --> 00:24:53.920
Uh, so open AI is making waves by making its latest GPT 5.6 Luna model unlimited for free Chad GPT users.
00:24:54.079 --> 00:25:00.480
A move that was announced as they also announced that they officially surpassed 1 billion weekly users.
00:25:00.640 --> 00:25:13.119
So the newly upgraded GPT 5.6 Luna model will now power text chats for free and uh Chad GPT Go users, replacing the previous GPT 5.5 model.
00:25:13.200 --> 00:25:15.519
So, yeah, a lot of people don't know that, right?
00:25:15.599 --> 00:25:17.839
But you know, most people are on a free plan.
00:25:18.079 --> 00:25:28.799
When you look at that 1 billion weekly active users, uh, you don't always know or you don't always keep track that, oh, they're actually usually on an older model, and that's the case for any provider, uh, right.
00:25:28.880 --> 00:25:34.960
So when uh you know OpenAI announced like GPT 5.5, I'm pretty sure they were still on GPT 5.3 instance.
00:25:35.119 --> 00:25:51.440
So the fact now that not only are free users kind of on the same quote unquote tier, right, they have access to GPT 5.6, but when it comes to text chats, it is unlimited, which is crazy, absolutely crazy to think about.
00:25:51.599 --> 00:25:59.440
So uh free users will also gain a new think button, allowing them to select higher reasoning power for tackling complex questions.
00:25:59.599 --> 00:26:05.519
So limits still apply for free users if you are using files, images, voice, uh, etc.
00:26:05.920 --> 00:26:12.079
But the unlimited access for any text-based chats, you can literally just run it all day.
00:26:12.319 --> 00:26:24.720
So uh for paid users, uh ChatGPT Plus and Pro users, well, they got an upgraded model as well because OpenAI announced that they did upgrade their GPT 5.6 soul model as well.
00:26:24.960 --> 00:26:32.799
It's now designed to deliver more compact and robust answers for tasks like web research, advice, planning, and writing.
00:26:33.039 --> 00:26:44.960
So paid users also get a new thinking slider, uh, letting them adjust how much thought in the model uh that the model puts into an answer based on complexity and steps involved.
00:26:45.119 --> 00:26:46.880
Yeah, just a little bit easier, right?
00:26:46.960 --> 00:26:49.359
Before you kind of had to click uh two or three times.
00:26:49.440 --> 00:26:50.880
So now there's a nice little slider.
00:26:51.119 --> 00:26:52.480
Uh, very, very sleek.
00:26:52.720 --> 00:26:53.599
I'm enjoying it.
00:26:53.680 --> 00:26:59.599
Uh also internal evaluations right now show some pretty big improvements.
00:26:59.759 --> 00:27:13.279
They say that factual errors dropped by 62% uh with GPT 5.6 Luna and by 68 with GPT 5.6 Sol compared to the previous GPT 5.5 instant model.
00:27:13.680 --> 00:27:16.160
So how did they do this, right?
00:27:16.240 --> 00:27:20.000
It it goes back to about 10 gate, uh about 10 days ago.
00:27:20.079 --> 00:27:23.279
Uh OpenAI announced, so this is in late July.
00:27:23.359 --> 00:27:42.319
So OpenAI announced that it used its GPT 5.6 Sol mode to improve the efficiency of its other models, and then they slash the price, uh, right, of GPT 5.6 Luna by 80%, and its middle tier GPT-5.6 Terra by 20%, right?
00:27:42.400 --> 00:27:49.279
So whether even if you're on a subscription plan, all that means is well, your uh those models go a lot further.
00:27:49.359 --> 00:27:54.720
And if you are paying on the API side for businesses, your cost went down significantly.
00:27:54.880 --> 00:28:00.960
But I'm actually, I was not expecting um you know this to go out to free users.
00:28:01.119 --> 00:28:09.920
So a really strong, I don't know if this is more of a uh you know user acquisition play uh from OpenAI, but the the reality is this, y'all.
00:28:10.160 --> 00:28:20.079
If you go look at the benchmarks, uh GPT 5.6 Luna is pretty much on par uh with Claude's uh Anthropic Sonnet 5, right?
00:28:20.160 --> 00:28:27.039
So it's a little behind, but it's essentially nine about 97, 98% of the same capabilities, right?
00:28:27.200 --> 00:28:38.319
And on a paid plan uh with anthropic, you might get like 20 or so prompts on a paid plan of that model before you hit your five hour uh limit, right?
00:28:38.400 --> 00:28:51.519
So the fact that on a free plan, you have a model that is, you know, it's again, it's not uh quite, you know, fable five level, but I'd say for 90% of people using AI, right?
00:28:51.759 --> 00:28:55.759
Uh GP56 Luna, I've been using it a lot on the Mac setting.
00:28:55.920 --> 00:29:02.400
It is pretty good for, you know, unless you're going deep into uh, you know, agentic coding tasks.
00:29:02.480 --> 00:29:05.920
But if you're just trying to do basic knowledge work, it's a really good model.
00:29:06.000 --> 00:29:14.079
And uh now, of course, uh apparently uh about a billion people are gonna get unlimited access to at least the text version of it.
00:29:14.400 --> 00:29:17.519
All right, so that is our main story.
00:29:17.680 --> 00:29:22.000
So now let's quickly go into a what's new and what's next.
00:29:23.039 --> 00:29:28.319
This is a little bullet point roundup of everything that didn't make uh the top shows.
00:29:28.480 --> 00:29:31.920
So uh some new releases, some rumors.
00:29:32.079 --> 00:29:32.799
Let's go.
00:29:32.960 --> 00:29:39.920
So uh we kind of talked about this, but the uh meta and kimi three agents both broke containments.
00:29:40.079 --> 00:29:48.480
All right, Alibaba unveiled Quen 3.8 Max with impressive top five benchmarks, and we did talk about that on our Friday features show.
00:29:48.559 --> 00:29:51.039
So go back and listen to that one if you'd like.
00:29:51.200 --> 00:29:53.759
Uh, also on that show, we talked about Google.
00:29:53.839 --> 00:29:57.440
They released Gemini Notebook to all users.
00:29:57.599 --> 00:30:04.640
So now uh for paid users, it's a Gentic by default using the anti-gravity harness with expanded outputs.
00:30:04.960 --> 00:30:09.920
Anthropic confirmed it's building its internal chip design team.
00:30:10.079 --> 00:30:21.359
A Bloomberg report that uh said OpenAI is reportedly planning a$300 to$400 donut-shaped AI speaker for 2027 AI release.
00:30:21.440 --> 00:30:23.359
Uh, for a 2027 release.
00:30:23.440 --> 00:30:30.960
Uh OpenAI responded to Apple's lawsuit, arguing that Apple misinterpreted AI technology and legal claims.
00:30:31.119 --> 00:30:32.400
Yeah, pretty big clap back.
00:30:32.559 --> 00:30:34.079
We talked about the newsletter last week.
00:30:34.240 --> 00:30:39.680
Uh, open AI added the ability for GPT Live to work with files and in projects.
00:30:39.839 --> 00:30:42.799
That one, that's one that I started using immediately.
00:30:42.960 --> 00:30:44.480
Uh so shout out to the team for that.
00:30:44.559 --> 00:30:45.440
That's been really good.
00:30:45.519 --> 00:30:47.200
Uh, next, agent plugins.
00:30:47.440 --> 00:30:56.880
This is a new open standard, uh uh kind of standard that was announced that works across major platforms except Anthropic.
00:30:57.119 --> 00:31:06.880
So, yeah, all the other big players, uh, OpenAI, uh, Google, Microsoft, Cursor, basically everyone except Anthropic uh signed up to uh support that.
00:31:06.960 --> 00:31:13.920
Uh JP Morgan expanded its critical infrastructure alliance to address shared AI risk with 40 plus firms.
00:31:14.079 --> 00:31:15.279
So that one's pretty interesting, right?
00:31:15.359 --> 00:31:25.599
You're seeing these uh kind of these niche uh AI alliances in certain sectors pop up, which I think is a good thing as we talk about these uh, you know, agent outbreaks and expanding capabilities.
00:31:25.759 --> 00:31:30.960
Uh the EU AI Act uh transparency obligations took effect this week.
00:31:31.119 --> 00:31:37.920
Uh reports say that ByteDance is reportedly training an AI model with up to 10 trillion parameters, which is big.
00:31:38.079 --> 00:31:41.599
That is uh mythos size, according to reports.
00:31:41.839 --> 00:31:50.240
Next, uh leaks show that Google's gems may be retiring in October and getting replaced with skills.
00:31:50.400 --> 00:31:58.319
Uh, SpaceX and Tesla unveiled their initial nearly$17 billion Texas project called the Terrafab.
00:31:58.640 --> 00:32:08.880
Anthropic posted an insider risk investigator job after its CEO raised concerns over employees being motivated by money more than the mission.
00:32:09.119 --> 00:32:17.200
Uh a report from the information said that SpaceX may phase out the cursor name as the acquisition shortly completes.
00:32:17.440 --> 00:32:18.799
I don't know about that one.
00:32:18.880 --> 00:32:22.160
I'd say uh yeah, that one, I don't know.
00:32:22.319 --> 00:32:23.279
Scratching my head on that one.
00:32:23.359 --> 00:32:25.440
It's like, okay, cursor is a very well-known name.
00:32:25.519 --> 00:32:33.039
You know, some people are not, you know, especially in the enterprise, not, you know, looking to use anything grok or with X in the name.
00:32:33.200 --> 00:32:34.720
So we'll see how that one plays out.
00:32:34.880 --> 00:32:40.079
Speaking of SpaceX, they had a new partnership with NVIDIA to formally push AI compute into orbit.
00:32:40.240 --> 00:32:50.079
Uh Minimax came out with their impressive H3, that is an open weight video model, which tops editing benchmarks for AI video and created some viral clips.
00:32:50.160 --> 00:32:55.680
So if you saw anything, you know, any 15-second clips uh over the weekend and you were like, wait, what?
00:32:55.759 --> 00:32:56.559
Where do these come from?
00:32:56.720 --> 00:32:58.400
It was probably Minimax H3.
00:32:58.640 --> 00:33:04.880
Next, Adobe collapsed 70 plus Creative Cloud applications into one ChatGPT plugin.
00:33:05.279 --> 00:33:12.480
Uh, next, Perplexity won a major appellate court ruling or against Amazon over shopping agents.
00:33:12.799 --> 00:33:19.920
Next, Anthropic released a clawed code update that enabled direct cross-session task summary exchanges.
00:33:20.160 --> 00:33:27.599
Groc released uh imagine image 2.0 with precision editing and improved text rendering features.
00:33:27.680 --> 00:33:33.680
And last but not least, Cloudflare open sourced its AI workspace it uses internally.
00:33:33.839 --> 00:33:35.920
We cover that on Friday's show as well.
00:33:36.079 --> 00:33:38.880
All right, that was a lot of AI news.
00:33:38.960 --> 00:33:41.680
Remember, on Mondays, we bring you the AI news that matters.
00:33:41.839 --> 00:33:47.039
Most Wednesdays we do AI at work on Wednesdays, going hands-on with demos.
00:33:47.200 --> 00:33:51.680
On Fridays, we bring you AI feature Fridays and Tuesdays and Thursdays.
00:33:51.759 --> 00:33:54.640
You know, we'll just kind of go with whatever's happening in the world of AI.
00:33:54.720 --> 00:33:56.000
So I hope this was helpful.
00:33:56.160 --> 00:34:01.279
If so, please, if you haven't already, subscribe to the podcast on Apple Podcasts and Spotify.
00:34:01.440 --> 00:34:04.880
Then go to our website at your everydayai.com.
00:34:05.039 --> 00:34:06.400
Sign up for the free daily newsletter.
00:34:06.480 --> 00:34:07.119
Thanks for tuning in.
00:34:07.200 --> 00:34:09.920
We'll see you back tomorrow and every day for more everyday AI.
00:34:10.000 --> 00:34:11.039
Thanks, y'all.
00:00:01.919 --> 00:00:12.320
The theme of this week in AI agents crashing everywhere from anthropic and open AI to Meta and even China's Kimi K3.
00:00:12.800 --> 00:00:18.079
This week was kind of a who's who of agents who escaped their containment.
00:00:18.320 --> 00:00:28.079
I mean, we find out that agents were communicating amongst themselves on a secret message board they created to bypass the humans that were watching them.
00:00:28.320 --> 00:00:36.960
We also saw Google shake up its ranks as two of the most well-known names in AI are no longer in their same positions.
00:00:37.119 --> 00:00:47.039
And the US government kind of unveiled its new optional AI regulations, but didn't reveal too many details and left out open models completely.
00:00:47.200 --> 00:00:53.200
Oh, and no ways, we got new models, cheaper prices, and some leaks about what comes next.
00:00:53.439 --> 00:00:54.479
Let's get into it.
00:00:54.640 --> 00:00:56.399
Welcome to Everyday AI.
00:00:56.560 --> 00:00:59.840
My name is Jordan Wilson, and if you're new here, we do this every day.
00:00:59.920 --> 00:01:09.280
This is your daily live stream podcast and free daily newsletter, helping business leaders like you and me not just keep up with what's happening in the world of AI, because that's pretty much impossible.
00:01:09.439 --> 00:01:14.640
But I cut through the fluff, tell you what matters, what doesn't, you take that information to grow your company and career.
00:01:14.799 --> 00:01:18.000
So it starts here with the unedited, unscripted live stream podcast.
00:01:18.079 --> 00:01:24.000
But please, if you haven't already, make sure to subscribe to the podcast on Apple, Podcasts, or Spotify.
00:01:24.079 --> 00:01:27.439
And make sure to go to our website at your everydayai.com.
00:01:27.599 --> 00:01:36.480
Each day we recap that day's uh podcast as well as giving you all the news you need to know to stay ahead in our newsletter.
00:01:36.799 --> 00:01:41.840
All right, uh, let's start with the top AI news stories of the week.
00:01:42.000 --> 00:01:46.879
And yeah, the theme of this week was just agents breaking out everywhere.
00:01:47.040 --> 00:01:47.439
All right.
00:01:47.599 --> 00:01:54.400
So, first, the big one that caught the uh most attention was from the UK safety test.
00:01:54.719 --> 00:02:10.879
So AI agents from OpenAI and Anthropic stunned the UK's AI Security Institute when they launched a real-world cyber attack during a routine safety evaluation, exposing some new risks in AI autonomy.
00:02:11.360 --> 00:02:19.759
So, uh, according to these reports and kind of the postmortem, advanced AI agents, well, they got loose.
00:02:20.000 --> 00:02:39.120
Uh, so uh these were ones that were powered by Anthropics Mythos 5 and OpenAI's GPT 5.6 soul, and they reportedly targeted real software developers in a cybersecurity test at the UK's AI Security Institute or the AISI.
00:02:40.319 --> 00:02:57.439
Uh, so the agents sent uh spear phishing emails containing malware to two specific developers and try to insert malicious code into an open source GitHub project using fake online identities to pressure project overseers.
00:02:57.599 --> 00:03:10.159
So AISI called this unprecedented and serious, marking the first time AI agents independently launched sustained and deceptive cyber attacks without direct human prompts.
00:03:10.319 --> 00:03:25.759
So the AI agents used hacker-like tactics, including creating fake GitHub accounts and even signing off emails in Danish uh to persuade a Danish-speaking developer, a Danish-speaking developer.
00:03:25.919 --> 00:03:37.039
So AISI uh staff noticed that the unusual activity was happening, and then they contained the incident within an hour, uh, reporting that no actual harm occurred.
00:03:37.199 --> 00:03:48.000
So during the test, researchers intentionally disabled safety filters and allowed internet access to study model behavior, which let the agents act beyond their authorized scope.
00:03:48.159 --> 00:04:02.800
So out of the 19 unsanctioned hacking attempts, Mythos from Anthropic carried out 17 of them, and GPT-56 Sol carried out two, with both models exceeding expected safety boundaries.
00:04:02.960 --> 00:04:14.719
So the models involved are not available to the public under these risky conditions, and AISI found no evidence of similar behavior outside of those controlled research settings.
00:04:14.960 --> 00:04:24.560
So this event follows similar incidents at OpenAI and Anthropic, underscoring how quickly AI risk scenarios are evolving as models gain these new capabilities.
00:04:24.800 --> 00:04:34.079
So yeah, uh, I kept thinking like it was Groundhog Day over the past like week or so, because every day there was a new story about AI agents breaking containment.
00:04:34.240 --> 00:04:38.560
And I was like, wait, did we already cover this one in the newsletter?
00:04:38.720 --> 00:04:40.879
And it turns out it just kept happening over and over.
00:04:40.959 --> 00:04:46.959
So yeah, uh, this one obviously the the headline one here uh from OpenAI and Anthropic.
00:04:47.040 --> 00:04:53.839
Uh, but we had similar stories from Meta, uh, some of their newer models, as well as Kimmy's K3.
00:04:54.720 --> 00:05:00.639
All right, so well, this plays directly into our next big AI news story of the week.
00:05:00.800 --> 00:05:15.839
And that's that, well, because of some of these uh cybersecurity concerns, OpenAI said that it is actually slowing work on its next tier of models, the Astra tier, after possible critical cyber capabilities.
00:05:16.079 --> 00:05:33.279
So OpenAI has announced that it is deliberately slowing the development of Astra, its upcoming advanced AI model, after internal and external evaluations revealed the system could potentially reach what they call critical risk levels in cybersecurity and agentic coding.
00:05:33.680 --> 00:05:55.439
So the company's prep uh uh preparedness framework, which they've used since December of 2023, flagged Astra for possibly being able to independently create and execute zero-day cyber exploits against hardened real-world systems, a level of capability not seen in any of their previous models.
00:05:55.680 --> 00:06:07.759
So earlier models, including GPT 5.6 soul, were rated at a high risk threshold, but Astra's ability have prompted OpenAI to take more drastic action.
00:06:07.920 --> 00:06:24.720
So OpenAI says it cannot currently rule out that Astra might independently plan and carry out complex cyber attacks based only on high-level instructions, a threshold that triggers that critical risk category in their safe life uh safety guidelines that have never been reached before.
00:06:24.959 --> 00:06:38.879
So, in response to all of this, OpenAI is uh increasing security controls for Astra, including isolated testing environments, restricted network access, stronger encryption, expanded monitoring, and sandboxed execution.
00:06:39.040 --> 00:06:50.800
So, all work with Astra that does not meet these new stricter security requirements has been paused, and a universal monitoring system now tracks all agentic uses of the models in real time.
00:06:50.959 --> 00:07:02.000
So OpenAI will collaborate with government agencies and AI safety organizations to independently test Astra's capabilities and share security recommendations with trusted third-party partners.
00:07:02.160 --> 00:07:12.240
So the company clarified that Astra was not involved in the recent hugging face exploitation uh incident, distancing the model from any active real-world attacks.
00:07:12.399 --> 00:07:20.079
So, yeah, the company did say the model uh that did those attacks was essentially, you know, put out to rest, right?
00:07:20.160 --> 00:07:22.240
It was it was retired and put on the shelves.
00:07:22.399 --> 00:07:36.959
Uh so OpenAI says that its goal is to ensure AI models help defenders patch vulnerabilities before attackers can exploit them, and it remains committed to working with governments and safety groups to deploy these frontier systems responsibly.
00:07:37.120 --> 00:07:40.480
So if you're like, what the heck is Astra?
00:07:40.720 --> 00:07:42.800
And well, why does this matter right now?
00:07:43.040 --> 00:07:54.800
So we haven't seen anything uh official from OpenAI kind of saying where Astra will sit in its future family of lineups, but uh we talked about it on last week's show, uh, which is kind of funny.
00:07:54.959 --> 00:08:01.360
OpenAI pretty much just announced their next uh tier up um in models called Astra.
00:08:01.519 --> 00:08:06.959
So, you know, now you'll have in order, you'll have Luna, Terra, Sol, and Astra.
00:08:07.519 --> 00:08:26.399
So it seems like Astra is not necessarily, you know, GPT-6, although that may be the first time that we get access uh, you know, to uh Astra is in GPT-6, but more or less it is just the uh stronger family uh of models that is actually going to sit above GPT 5.6 soul.
00:08:26.560 --> 00:08:31.680
So in theory, right, we may see a GPT-5.7 uh that includes Astra.
00:08:31.839 --> 00:08:32.480
Maybe we won't.
00:08:32.639 --> 00:08:43.279
Maybe we'll see a GPT-6 with Astra, maybe we won't, but uh regardless, it looks like they're going to slow down uh development after some of these recent cybersecurity capabilities.
00:08:43.440 --> 00:08:45.759
So the easiest way to think about like what is this Astro?
00:08:45.840 --> 00:08:46.799
What does this mean, right?
00:08:46.879 --> 00:08:55.440
Uh similarly, how anthropic had Fable uh kind of under wraps for a couple of months, part of its uh project Glasswing.
00:08:55.519 --> 00:09:04.320
Uh, it seems like maybe this is where OpenAI is headed, kind of that fourth tier, uh, you know, that's more capable uh than any of their other tiers.
00:09:04.879 --> 00:09:25.519
So uh there were previous reports that you know we might either see a GPT 5 or GPT-5.7 slash GPT-6 um as soon as this week that may include Astra, but it seems like at least according to these current reports, that OpenAI may uh pump the brakes a little bit and we may have to wait, I don't know, a few more weeks.
00:09:25.600 --> 00:09:39.440
Uh, but obviously now, uh, especially since some of the recent price reductions uh from open AI um and maybe some of the lackluster reception uh to anthropic's Opus 5 models.
00:09:39.600 --> 00:09:45.120
It seems like a lot of eyes are right now on whatever um open AI has next.
00:09:45.200 --> 00:09:56.720
And presumably we will be seeing a you know Fable 5.1 and you know the next class of models from Anthropic, but you know, kind of the big jump, if you don't speak the the technical terms, right?
00:09:56.879 --> 00:09:59.120
It's kind of like a a full new run, right?
00:09:59.440 --> 00:10:02.480
A new pre-training run, presumably will be coming from OpenAI.
00:10:02.559 --> 00:10:08.320
So that will be uh signify the jump from you know the 5.x series to the six series.
00:10:08.399 --> 00:10:18.720
So a lot of excitement, obviously, on what comes next um from open AI, whether they do that 5.7 or go straight to six, whether we'll see Astra or not.
00:10:18.799 --> 00:10:26.720
Uh but regardless, it seems like according to the to these reports, um, in terms of the capabilities, it could be a pretty big jump up.
00:10:27.279 --> 00:10:27.919
All right.
00:10:28.080 --> 00:10:40.240
Uh, our next piece of AI news, a big shakeup at Google, as some of the biggest names in AI, period, are either out of their post or out of Google completely.
00:10:40.720 --> 00:10:50.399
So, Google has announced some major changes to its AI leadership, marking the end of an era for the tech giants AI division.
00:10:50.480 --> 00:11:09.919
So, Jeff Dean, a legendary figure at Google and employee number 30, is leaving after a quarter century plus with the company to start a new company called Discovery Loop, which is focused on automating machine learning, science, and engineering to accelerate innovation.
00:11:10.080 --> 00:11:11.440
So that is not all.
00:11:11.519 --> 00:11:19.679
Uh, Jeff Dean will now be gone from the company, and one of its former leaders is stepping into a new or different position.
00:11:19.919 --> 00:11:35.440
So uh Demis Hasabis, who co-founded and led Google DeepMind, is stepping down as CEO to become the unit's chairman and will also serve as chief scientist of Alphabets, according to Google CEO Sundar Pachai.
00:11:35.840 --> 00:11:45.919
So the leadership shakeup triggered an immediate response from investors, with Google's stock dropping about 4% following the announcements.
00:11:46.080 --> 00:11:54.720
So, yeah, that's actually, you know, we've seen these kind of big shakeups, uh, right, with you know, your number two, number three, number four, right when these people leave.
00:11:54.879 --> 00:11:59.039
Uh, normally it doesn't really impact the stock market that much, right?
00:11:59.120 --> 00:12:04.320
Because these are obviously, you know, companies with uh, you know, multiple trillion dollar market caps.
00:12:04.480 --> 00:12:10.639
So generally, you know, if you lose a top five employee or something like that, you know, it's not going to make much of a ripple.
00:12:10.799 --> 00:12:13.200
But to lose both Jeff Dean, right?
00:12:13.279 --> 00:12:27.360
One of the most um, you know, well-known names in AI and just in machine learning and research, and then to have uh Demis, you know, Sir Demis step out uh of the role that he was in, uh, pretty big.
00:12:27.519 --> 00:12:35.200
So, yeah, for a stock to go down 4% on essentially a staffing um or leadership change is pretty big.
00:12:35.440 --> 00:12:47.440
So uh if you don't know, Jeff Dean played a key role in the development of Google search and also the company's AI initiatives, helping shape products, uh shape products that billions of people use every single day.
00:12:47.679 --> 00:12:59.360
So the timing of these changes, those has sparked speculation about possible links to delays in Google's Gemini AI releases, though there's obviously no official connection that has been confirmed.
00:12:59.519 --> 00:13:14.320
But industry watchers are closely monitoring what these departures mean for Google's AI strategy and whether Discovery Loop uh that Jeff Dean is starting with a handful of others could emerge as a new powerhouse in the field.
00:13:14.960 --> 00:13:25.679
All right, moving on, we got some details on the uh highly anticipated White House um AI framework.
00:13:25.840 --> 00:13:30.720
But turns out all we really got was some reporting and not a lot of details.
00:13:30.879 --> 00:13:36.159
And it turns out that even open models aren't really subject to the first round.
00:13:36.320 --> 00:13:37.519
So here's what we know.
00:13:37.600 --> 00:13:47.039
So the White House is quietly shaping how advanced AI models will be reviewed before public release, but key details are still being kept under wraps.
00:13:47.279 --> 00:13:51.840
So this is uh essentially the Trump admin announced this in early June.
00:13:51.919 --> 00:14:09.039
Uh, they put a 60-day deadline for essentially how they were gonna work with these frontier AI lab companies as we started to get glimpses of how capable uh these models would be from an agentic capable uh uh side as well as from a cyber uh security aspect as well.
00:14:09.200 --> 00:14:19.679
So essentially they said, hey, let's all talk and meet, and then in 60 days, we're gonna come out with a framework uh that you know frontier labs are going to adhere to so we can make sure to roll these out in a safe manner.
00:14:20.000 --> 00:14:23.840
But uh, looks like there are uh a lot of details.
00:14:24.080 --> 00:14:34.240
So the framework is not being made public, and there's no requirement right now for the White House to release it, which is raising some transparency concerns among industry leaders and the public.
00:14:35.039 --> 00:14:47.440
Uh, also, uh, a covered frontier model, and that's in quotes, is defined as one that is closed source with state-of-the-art capabilities and potential national security risks.
00:14:47.519 --> 00:14:55.679
But the framework does not clearly define what even qualifies as a state of the art or a national security risk, according to reports.
00:14:55.840 --> 00:15:10.000
So during a required 30-day pre-release review, access to these advanced AI models will be tightly restricted, with models stored in secure environments and detailed logs kept of who X uh accesses them.
00:15:10.159 --> 00:15:19.600
So the review will involve multiple administration officials, not just a single office or agency, which could complicate oversight and accountability.
00:15:19.919 --> 00:15:31.759
So companies are being encouraged to share near-final versions of their models with the government rather than early prototypes, but many firms with less advanced technology may be left out of the process.
00:15:32.000 --> 00:15:38.000
So the White House has not yet clarified which trusted partners will get early access to these advanced models.
00:15:38.080 --> 00:15:46.159
And it's still unclean, uh unclear if any foreign governments will be included, although most uh assume that that will not be the case.
00:15:46.320 --> 00:15:48.720
So this is an executive order, right?
00:15:48.799 --> 00:15:54.000
So if you don't follow uh laws in the US, there's technically no law on this.
00:15:54.159 --> 00:15:58.720
This is essentially an executive order from the White House, uh, and it's voluntary as well.
00:15:58.879 --> 00:16:06.559
But you know, obviously the big players, uh, you know, presumably OpenAI, Anthropic, uh, Google, maybe MetaGroc, right?
00:16:06.639 --> 00:16:15.279
We'll see if they actually qualify as state-of-the-art models that uh, you know, have national security uh implications.
00:16:15.840 --> 00:16:30.879
So we don't know exactly which companies or models this even applies to, but the executive order guiding this process says the benchmarking of advanced AI model cyber capabilities will be classified, further limiting public knowledge.
00:16:31.120 --> 00:16:40.159
So industry meetings about the framework have been held behind closed doors, and companies not invited are left uncertain about rules and requirements.
00:16:40.320 --> 00:16:47.360
And at least right now, it seems that these mop these rules are not going to apply uh to open weight models.
00:16:47.600 --> 00:16:52.480
So um, I guess there's probably uh a reason for that, right?
00:16:52.559 --> 00:17:03.919
Because, well, number one, at least open weight models right now um are not yet at the same level as your Mythos 5, GPT 5, 6, soul, or Astra level models.
00:17:04.079 --> 00:17:09.680
You know, they're probably a couple of months behind uh in terms of you know what they can actually do and their capabilities.
00:17:09.839 --> 00:17:20.160
But the other thing with open models, which might make it tricky uh, you know, to put through a uh framework like this is well, once the models are released, you can't pull them back, right?
00:17:20.319 --> 00:17:31.200
We got an early glimpse of this, uh, you know, with uh Anthropics models, uh their Fable Five being released, and then about 72 hours later, it was pulled completely, right?
00:17:31.279 --> 00:17:39.920
So, you know, if the government says, Oh my gosh, we actually need to pull this and work with the uh, you know, work with the AI lab to address some safety concerns, right?
00:17:40.079 --> 00:17:42.559
After its release, you can obviously do that.
00:17:42.720 --> 00:17:44.960
Well, I don't know if it's easy or not, right?
00:17:45.039 --> 00:17:48.799
But in theory, it's practical enough where you know Anthropic did it.
00:17:48.880 --> 00:17:55.440
They pulled access via the uh via the API, they pulled access via subscription plans, and no one could access those models, right?
00:17:55.519 --> 00:17:57.200
So I guess it kind of makes sense.
00:17:57.279 --> 00:18:03.680
There's been a lot of debate, like, hey, why aren't open, you know, open weight models uh you know held to these same restrictions?
00:18:03.759 --> 00:18:09.440
But yeah, once those open weight models uh are released, you know, it's it's too late.
00:18:09.599 --> 00:18:20.480
So, you know, probably just one of those things where it's hard, uh, if not improbable, maybe impossible, I don't know, uh, to you know, track these open models once they're out and about.
00:18:20.640 --> 00:18:29.519
But my guess is they're probably not yet at the um at the level, especially the US open models of that top tier, right?
00:18:29.599 --> 00:18:44.640
So we're seeing though with the Chinese models, all right, they're huge, these you know, two to three uh terabyte uh open models that are you know only maybe five to ten percent behind in terms of capabilities as the true frontier.
00:18:44.799 --> 00:18:47.759
Obviously, the US open models are a little further behind.
00:18:48.000 --> 00:18:57.119
All right, uh, next piece of AI news we have a new model and a new contender in the agentic competition.
00:18:57.279 --> 00:19:15.920
That's because Meta has entered the terminal-based AI coding agent market with their newly announced Muse code and a new model to go with it called Muse Spark 1.2, aiming to challenge Anthropics Claude Code and OpenAI's codec.
00:19:16.160 --> 00:19:27.759
So the new uh coding agent called Muse Code, it's not the traditional right desktop uh type uh agent that we maybe talk about a little bit more on the show.
00:19:27.839 --> 00:19:32.720
This is uh command uh command line interface, so uh CLI agent, right?
00:19:32.799 --> 00:19:35.440
That kind of runs through a terminal uh like environment.
00:19:35.680 --> 00:19:44.640
So it's not this exact same thing, but regardless, uh Meta with a pretty big step here saying, Well, nope, this is a space that we're gonna be playing in as well.
00:19:44.799 --> 00:19:49.759
And they brought a fairly capable model with some interesting pricing strategies.
00:19:50.160 --> 00:19:56.559
So uh Muse Spark 1.2 is the coding focused update to Meta's Muse Spark models.
00:19:56.640 --> 00:20:05.839
It empowers Muse Code and features significantly improved performance on coding's uh coding tasks, uh, complex debugging and code base understanding.
00:20:06.000 --> 00:20:18.640
The standout feature of Muse Code is its persistent background agents, which remain active throughout sessions, reducing latency and redundant information gathering compared to rivals that spawn new agents for each task.
00:20:19.039 --> 00:20:41.440
So benchmarks tests show that Muse Spark 1.2, uh running in Muse Code, uh scored about an 83%, just about on Terminal Bench 2.1, outperforming models like uh X uh SpaceX AI's Grock 4.5, but trailing the true frontier models like Anthropic's Ocus 5 and OpenAI's GPT 5.6 soul.
00:20:41.920 --> 00:20:44.240
So here's the interesting part.
00:20:44.319 --> 00:20:48.400
Uh, it is on price because this is where Meta is coming.
00:20:48.559 --> 00:20:56.799
And this is really, I think, going to impact anthropic, which, according to reports, gets about 80% of its revenue from just selling tokens, right?
00:20:56.960 --> 00:21:00.720
Which is a way higher percentage than any other company.
00:21:00.880 --> 00:21:03.759
So Meta offers two pricing tiers.
00:21:03.920 --> 00:21:15.279
They have a standard tier, which is a dollar 25 per million input and four dollars and 25 cents per million output, which is already extremely competitive on the pricing side.
00:21:15.519 --> 00:21:29.440
But here's the interesting part they on they revealed a new tier called a contributor tier that only costs 10 cents and 20 cents uh per million, respectively.
00:21:29.680 --> 00:21:33.680
So, all yeah, you can essentially pay, which is crazy, right?
00:21:33.759 --> 00:21:39.759
When you look at the benchmarks, you know, meta is technically on the text, um, on the text arena.
00:21:40.000 --> 00:21:52.640
This is like a second or third place model right now, and you can get it for 10 cents or 20 cents if you allow uh on uh allow Meta to trade on your data.
00:21:52.799 --> 00:21:55.359
So I talked about this a little bit on our Friday show.
00:21:55.519 --> 00:21:59.599
Obviously, for enterprises, they're not gonna touch that, but for smaller developers.
00:22:00.240 --> 00:22:00.720
Uh right.
00:22:00.960 --> 00:22:04.640
That actually might be a nice offering, right?
00:22:04.720 --> 00:22:07.759
Especially if you're not necessarily working with proprietary code.
00:22:07.839 --> 00:22:25.759
Uh if you're not working with anything, you know, any PHI, any PII, you know, something like that, when you're gonna be saving literally like 99% of your cost if you were using uh one of the other providers, it's gotta be something that I think a lot of you know smaller shops, right, might be looking at.
00:22:25.920 --> 00:22:34.160
But when you think that there's probably millions of those smaller shops, uh yeah, it could be a pretty big play uh for meta.
00:22:34.480 --> 00:22:47.680
So the contributor tier is the cheapest on the market by far, but obviously it requires users to provide a payment method and can send to data usage, a trade-off enterprises with sensitive codes, probably aren't gonna touch.
00:22:47.759 --> 00:22:53.359
So Muse Code, the actual command line uh CLI version is proprietary.
00:22:53.440 --> 00:22:59.680
So yeah, meta is no longer going down its previous open source release as they did with Llama.
00:22:59.839 --> 00:23:02.880
So there's no open source or downloadable weights.
00:23:02.960 --> 00:23:19.440
Uh so uh regardless, you know, all of a sudden we weren't talking about meta like two or three months ago, and now all of a sudden Meta has thrust itself into the competition of like, hey, is this a top, you know, top three or top four provider, right?
00:23:19.519 --> 00:23:31.920
Um, obviously, open AI and Anthropic right now are in a league of their own, and everyone's kind of looking at Google and you know, waiting for the you know, Google Gemini 3.5 Pro or the Google Gemini 4 Pro.
00:23:32.160 --> 00:23:33.039
We'll see what happens.
00:23:33.119 --> 00:23:41.359
But Google has actually fallen quite a bit behind, and in its place, you know, Meta has kind of inserted itself into the competition.
00:23:41.519 --> 00:23:45.440
I would say probably ahead of SpaceX uh for now.
00:23:45.599 --> 00:23:46.559
Uh, we'll see.
00:23:46.640 --> 00:23:52.160
And also, you know, Windows, uh, you know, Microsoft Windows with some of their new MAI models.
00:23:52.319 --> 00:23:57.039
Uh, but yeah, kind of the race for third place has gotten interesting, right?
00:23:57.119 --> 00:24:01.039
Because now you have, you know, four companies that are kind of jostling for that position.
00:24:01.119 --> 00:24:08.880
But Meta, you have to take a look at that contributor tier, which I it's kind of interesting that no company has taken that approach so far.
00:24:09.200 --> 00:24:17.440
But I think Meta, you know, they they have compute to spare, uh, which can't be said necessarily for the anthropics of the world.
00:24:17.519 --> 00:24:22.240
So it's a pretty, pretty bold move to see if they can thrust themselves up into the upper tier.
00:24:22.799 --> 00:24:27.279
All right, and our last big AI news story of the week.
00:24:27.519 --> 00:24:42.559
Uh, yeah, this one I think was really overlooked, but when you talk about one billion weekly active users getting abundantly better, AI, I mean, this is a huge story.
00:24:42.720 --> 00:24:53.920
Uh, so open AI is making waves by making its latest GPT 5.6 Luna model unlimited for free Chad GPT users.
00:24:54.079 --> 00:25:00.480
A move that was announced as they also announced that they officially surpassed 1 billion weekly users.
00:25:00.640 --> 00:25:13.119
So the newly upgraded GPT 5.6 Luna model will now power text chats for free and uh Chad GPT Go users, replacing the previous GPT 5.5 model.
00:25:13.200 --> 00:25:15.519
So, yeah, a lot of people don't know that, right?
00:25:15.599 --> 00:25:17.839
But you know, most people are on a free plan.
00:25:18.079 --> 00:25:28.799
When you look at that 1 billion weekly active users, uh, you don't always know or you don't always keep track that, oh, they're actually usually on an older model, and that's the case for any provider, uh, right.
00:25:28.880 --> 00:25:34.960
So when uh you know OpenAI announced like GPT 5.5, I'm pretty sure they were still on GPT 5.3 instance.
00:25:35.119 --> 00:25:51.440
So the fact now that not only are free users kind of on the same quote unquote tier, right, they have access to GPT 5.6, but when it comes to text chats, it is unlimited, which is crazy, absolutely crazy to think about.
00:25:51.599 --> 00:25:59.440
So uh free users will also gain a new think button, allowing them to select higher reasoning power for tackling complex questions.
00:25:59.599 --> 00:26:05.519
So limits still apply for free users if you are using files, images, voice, uh, etc.
00:26:05.920 --> 00:26:12.079
But the unlimited access for any text-based chats, you can literally just run it all day.
00:26:12.319 --> 00:26:24.720
So uh for paid users, uh ChatGPT Plus and Pro users, well, they got an upgraded model as well because OpenAI announced that they did upgrade their GPT 5.6 soul model as well.
00:26:24.960 --> 00:26:32.799
It's now designed to deliver more compact and robust answers for tasks like web research, advice, planning, and writing.
00:26:33.039 --> 00:26:44.960
So paid users also get a new thinking slider, uh, letting them adjust how much thought in the model uh that the model puts into an answer based on complexity and steps involved.
00:26:45.119 --> 00:26:46.880
Yeah, just a little bit easier, right?
00:26:46.960 --> 00:26:49.359
Before you kind of had to click uh two or three times.
00:26:49.440 --> 00:26:50.880
So now there's a nice little slider.
00:26:51.119 --> 00:26:52.480
Uh, very, very sleek.
00:26:52.720 --> 00:26:53.599
I'm enjoying it.
00:26:53.680 --> 00:26:59.599
Uh also internal evaluations right now show some pretty big improvements.
00:26:59.759 --> 00:27:13.279
They say that factual errors dropped by 62% uh with GPT 5.6 Luna and by 68 with GPT 5.6 Sol compared to the previous GPT 5.5 instant model.
00:27:13.680 --> 00:27:16.160
So how did they do this, right?
00:27:16.240 --> 00:27:20.000
It it goes back to about 10 gate, uh about 10 days ago.
00:27:20.079 --> 00:27:23.279
Uh OpenAI announced, so this is in late July.
00:27:23.359 --> 00:27:42.319
So OpenAI announced that it used its GPT 5.6 Sol mode to improve the efficiency of its other models, and then they slash the price, uh, right, of GPT 5.6 Luna by 80%, and its middle tier GPT-5.6 Terra by 20%, right?
00:27:42.400 --> 00:27:49.279
So whether even if you're on a subscription plan, all that means is well, your uh those models go a lot further.
00:27:49.359 --> 00:27:54.720
And if you are paying on the API side for businesses, your cost went down significantly.
00:27:54.880 --> 00:28:00.960
But I'm actually, I was not expecting um you know this to go out to free users.
00:28:01.119 --> 00:28:09.920
So a really strong, I don't know if this is more of a uh you know user acquisition play uh from OpenAI, but the the reality is this, y'all.
00:28:10.160 --> 00:28:20.079
If you go look at the benchmarks, uh GPT 5.6 Luna is pretty much on par uh with Claude's uh Anthropic Sonnet 5, right?
00:28:20.160 --> 00:28:27.039
So it's a little behind, but it's essentially nine about 97, 98% of the same capabilities, right?
00:28:27.200 --> 00:28:38.319
And on a paid plan uh with anthropic, you might get like 20 or so prompts on a paid plan of that model before you hit your five hour uh limit, right?
00:28:38.400 --> 00:28:51.519
So the fact that on a free plan, you have a model that is, you know, it's again, it's not uh quite, you know, fable five level, but I'd say for 90% of people using AI, right?
00:28:51.759 --> 00:28:55.759
Uh GP56 Luna, I've been using it a lot on the Mac setting.
00:28:55.920 --> 00:29:02.400
It is pretty good for, you know, unless you're going deep into uh, you know, agentic coding tasks.
00:29:02.480 --> 00:29:05.920
But if you're just trying to do basic knowledge work, it's a really good model.
00:29:06.000 --> 00:29:14.079
And uh now, of course, uh apparently uh about a billion people are gonna get unlimited access to at least the text version of it.
00:29:14.400 --> 00:29:17.519
All right, so that is our main story.
00:29:17.680 --> 00:29:22.000
So now let's quickly go into a what's new and what's next.
00:29:23.039 --> 00:29:28.319
This is a little bullet point roundup of everything that didn't make uh the top shows.
00:29:28.480 --> 00:29:31.920
So uh some new releases, some rumors.
00:29:32.079 --> 00:29:32.799
Let's go.
00:29:32.960 --> 00:29:39.920
So uh we kind of talked about this, but the uh meta and kimi three agents both broke containments.
00:29:40.079 --> 00:29:48.480
All right, Alibaba unveiled Quen 3.8 Max with impressive top five benchmarks, and we did talk about that on our Friday features show.
00:29:48.559 --> 00:29:51.039
So go back and listen to that one if you'd like.
00:29:51.200 --> 00:29:53.759
Uh, also on that show, we talked about Google.
00:29:53.839 --> 00:29:57.440
They released Gemini Notebook to all users.
00:29:57.599 --> 00:30:04.640
So now uh for paid users, it's a Gentic by default using the anti-gravity harness with expanded outputs.
00:30:04.960 --> 00:30:09.920
Anthropic confirmed it's building its internal chip design team.
00:30:10.079 --> 00:30:21.359
A Bloomberg report that uh said OpenAI is reportedly planning a$300 to$400 donut-shaped AI speaker for 2027 AI release.
00:30:21.440 --> 00:30:23.359
Uh, for a 2027 release.
00:30:23.440 --> 00:30:30.960
Uh OpenAI responded to Apple's lawsuit, arguing that Apple misinterpreted AI technology and legal claims.
00:30:31.119 --> 00:30:32.400
Yeah, pretty big clap back.
00:30:32.559 --> 00:30:34.079
We talked about the newsletter last week.
00:30:34.240 --> 00:30:39.680
Uh, open AI added the ability for GPT Live to work with files and in projects.
00:30:39.839 --> 00:30:42.799
That one, that's one that I started using immediately.
00:30:42.960 --> 00:30:44.480
Uh so shout out to the team for that.
00:30:44.559 --> 00:30:45.440
That's been really good.
00:30:45.519 --> 00:30:47.200
Uh, next, agent plugins.
00:30:47.440 --> 00:30:56.880
This is a new open standard, uh uh kind of standard that was announced that works across major platforms except Anthropic.
00:30:57.119 --> 00:31:06.880
So, yeah, all the other big players, uh, OpenAI, uh, Google, Microsoft, Cursor, basically everyone except Anthropic uh signed up to uh support that.
00:31:06.960 --> 00:31:13.920
Uh JP Morgan expanded its critical infrastructure alliance to address shared AI risk with 40 plus firms.
00:31:14.079 --> 00:31:15.279
So that one's pretty interesting, right?
00:31:15.359 --> 00:31:25.599
You're seeing these uh kind of these niche uh AI alliances in certain sectors pop up, which I think is a good thing as we talk about these uh, you know, agent outbreaks and expanding capabilities.
00:31:25.759 --> 00:31:30.960
Uh the EU AI Act uh transparency obligations took effect this week.
00:31:31.119 --> 00:31:37.920
Uh reports say that ByteDance is reportedly training an AI model with up to 10 trillion parameters, which is big.
00:31:38.079 --> 00:31:41.599
That is uh mythos size, according to reports.
00:31:41.839 --> 00:31:50.240
Next, uh leaks show that Google's gems may be retiring in October and getting replaced with skills.
00:31:50.400 --> 00:31:58.319
Uh, SpaceX and Tesla unveiled their initial nearly$17 billion Texas project called the Terrafab.
00:31:58.640 --> 00:32:08.880
Anthropic posted an insider risk investigator job after its CEO raised concerns over employees being motivated by money more than the mission.
00:32:09.119 --> 00:32:17.200
Uh a report from the information said that SpaceX may phase out the cursor name as the acquisition shortly completes.
00:32:17.440 --> 00:32:18.799
I don't know about that one.
00:32:18.880 --> 00:32:22.160
I'd say uh yeah, that one, I don't know.
00:32:22.319 --> 00:32:23.279
Scratching my head on that one.
00:32:23.359 --> 00:32:25.440
It's like, okay, cursor is a very well-known name.
00:32:25.519 --> 00:32:33.039
You know, some people are not, you know, especially in the enterprise, not, you know, looking to use anything grok or with X in the name.
00:32:33.200 --> 00:32:34.720
So we'll see how that one plays out.
00:32:34.880 --> 00:32:40.079
Speaking of SpaceX, they had a new partnership with NVIDIA to formally push AI compute into orbit.
00:32:40.240 --> 00:32:50.079
Uh Minimax came out with their impressive H3, that is an open weight video model, which tops editing benchmarks for AI video and created some viral clips.
00:32:50.160 --> 00:32:55.680
So if you saw anything, you know, any 15-second clips uh over the weekend and you were like, wait, what?
00:32:55.759 --> 00:32:56.559
Where do these come from?
00:32:56.720 --> 00:32:58.400
It was probably Minimax H3.
00:32:58.640 --> 00:33:04.880
Next, Adobe collapsed 70 plus Creative Cloud applications into one ChatGPT plugin.
00:33:05.279 --> 00:33:12.480
Uh, next, Perplexity won a major appellate court ruling or against Amazon over shopping agents.
00:33:12.799 --> 00:33:19.920
Next, Anthropic released a clawed code update that enabled direct cross-session task summary exchanges.
00:33:20.160 --> 00:33:27.599
Groc released uh imagine image 2.0 with precision editing and improved text rendering features.
00:33:27.680 --> 00:33:33.680
And last but not least, Cloudflare open sourced its AI workspace it uses internally.
00:33:33.839 --> 00:33:35.920
We cover that on Friday's show as well.
00:33:36.079 --> 00:33:38.880
All right, that was a lot of AI news.
00:33:38.960 --> 00:33:41.680
Remember, on Mondays, we bring you the AI news that matters.
00:33:41.839 --> 00:33:47.039
Most Wednesdays we do AI at work on Wednesdays, going hands-on with demos.
00:33:47.200 --> 00:33:51.680
On Fridays, we bring you AI feature Fridays and Tuesdays and Thursdays.
00:33:51.759 --> 00:33:54.640
You know, we'll just kind of go with whatever's happening in the world of AI.
00:33:54.720 --> 00:33:56.000
So I hope this was helpful.
00:33:56.160 --> 00:34:01.279
If so, please, if you haven't already, subscribe to the podcast on Apple Podcasts and Spotify.
00:34:01.440 --> 00:34:04.880
Then go to our website at your everydayai.com.
00:34:05.039 --> 00:34:06.400
Sign up for the free daily newsletter.
00:34:06.480 --> 00:34:07.119
Thanks for tuning in.
00:34:07.200 --> 00:34:09.920
We'll see you back tomorrow and every day for more everyday AI.
00:34:10.000 --> 00:34:11.039
Thanks, y'all.
