Flight delayed 7 hours - asked Muse to file for compensation. 5 mins later i had $250 credit in my delta account. it even found and rebooked me a new flight.
it just figured everything out. even responded to the support email itself.
shit feels like magic.
so just to recap an insane day:
- anthropic broke the record for biggest IPO
- Google came out of nowhere to steal OpenAIs #2 spot, (gemini is back on top).
- GTA made $2 billion in 24 hours, the most money made by any piece of media ever.
- Chase bank got hacked, sensitive customer data stolen - likely by AI.
- NSA’s communication system is still down (3 days and counting)
- s&p500 is now down $6 trillion since peak
oh yeah and they’re a million AI-driven cyber attacks underway that everyone seems to be ignoring.
goodnight.
google is now shipping a new model every few days thats genuinely competitive to fable, GPT and opus
you don't achieve this without AI being heavily involved in the training of these models.
seems like they rebuilt the entire pre/post-training architecture and it paid off big time.
also notice how the recent models taste is much better and different to fable and gpt. also less throttled.
Taking a moment to reflect on the last 5 weeks of Gemini 4:
- Need more intelligence: we shipped Argon
- Need even more intelligence: we shipped Carbon (4 weeks later)
- Actually, we need it faster: we shipped Hydrogen (2 days later)
- Well, maybe more intelligence, just keep it fast: we're shipping Neon (6 days later)
We hear all the feedback and we're iterating as fast as we can. Keep telling us what's working (and what isn't) and let us know what you think of Neon!
Anyway, back to work..
BREAKING: Apple, $AAPL, is developing its own autonomous AI agent as it aims to rival Anthropic, OpenAI, and Meta, per Bloomberg.
The system will write code, browse the web, control apps, and navigate devices on a user’s behalf.
Shares of Meta, $META, and Anthropic, $ANT, are falling on the news.
🚨 Rumors I've been hearing about Google.
Google has closed the loop and unlocked RSI.
NEON
Gemini 4 Neon fixed all the problems people ever had with Gemini models. It already scores above Carbon and Argon on Artificial Analysis.
It is a SMALL model.
Internally Neon is s-rev28. The s means small. Hydrogen is s-rev27. Carbon and Argon are m-rev27. Google's own description said Neon "builds on Gemini 4 Hydrogen" at "Flash-tier speed and cost". Both lines are gone now.
REV28
Neon sits on the Hydrogen base. Google trained it further on rev28, a new post-training dataset, with a little sparkle of RSI on top. The knowledge cutoff moved to May 2026.
Carbon and Argon are still on rev27. The medium models have not gotten rev28 yet.
THE L MODELS
Sergey Brin said it in public a month ago. Google is "starting to get onto that RSI treadmill" and would be a couple of models ahead internally by the time anyone watched the video.
Small and medium are out. Google probably has L versions lying around right now doing the RSI work.
Four Gemini 4 models in seven weeks. Gemini moved more in the last month than OpenAI did in half a year.
Google will catch up. And destroy everyone.
just to be clear: google’s latest models aren’t even their best - they’re cooking up a much larger model (potentially several generations ahead):
- carbon and argon are considered “medium” models. neon is “small”
- yet these models outperform fable and gpt 6.2.
can you imagine the internal model?
google is now shipping a new model every few days thats genuinely competitive to fable, GPT and opus
you don't achieve this without AI being heavily involved in the training of these models.
seems like they rebuilt the entire pre/post-training architecture and it paid off big time.
also notice how the recent models taste is much better and different to fable and gpt. also less throttled.
these ai hacks won't stop until we have 24/7 defender models picking off the attackers.
anthropic's head of red-teaming said it best:
- every software platform will need to be re-built for the AI age.
- potentially rebuilt every single year.
its open season for anyone armed with the average open source model from 5 months ago.
‼️ BREAKING: Researchers broke past PHP's fix for the php2shell flaw and can again run their own code on web servers, no login needed.
Security firm depthfirst says the latest PHP releases are still vulnerable. The bug, CVE-2026-126714, is rated 9.2 out of 10.
https://t.co/EmNk1y4tjf
my thoughts on the anthropic ipo, the bull/bear case and the insane trajectory we're on to rsi.
whichever way you look at it: this is a historic event that'll have huge downstream effects on the industry
bookmark & give it a listen, lmk where you disagree
(p.s. apologies had to film this on my bed at tomorrowX lol)
my god anthropic IPO opening at $2 trillion which makes them the 6th largest company in the world. $500 billion in excess demand.
for anyone that thought the nvidia drama killed investor appetite its the opposite - money seems to be rotating heavily into the AI lab layer leaving semis to cool off for a bit.
phenomenal outcome for anthropic who were trying to launch at a 2T val.
things are going to get very wild.
BREAKING: Anthropic, $ANT, is now indicated to open at a $2 trillion market cap, up +33% from its IPO price.
This would mark a +57% increase from the bottom end of its initial IPO pricing range announced earlier this week.
Total demand for this IPO hit a massive $500 billion.
when are people going to wake up to strong possibility these hacks are coordinated by AI?
this morning Brussels government was hacked
yesterday 178 websites and platforms exploited.
the main tool being used is ARTEX - an open source AI penetration tester which people are strapping DeepSeek models to going to town on any nascent system.
governments, banks, web apps - anything is fair game.
i’m not one for doom mongering without strong evidence but on the contrary the amount of people i see here denying any possibility go AI being involved seems very artificial (pun intended)
We have learned that third parties who provide services to the firm were the target of a cyber attack.
Based on our investigation so far, we believe the attackers accessed certain marketing data held by those vendors, which may include some customer names and email addresses.
There is no disruption to our services.
Where appropriate, we will contact affected customers directly and offer support, including complimentary credit monitoring.
There is no connection to service disruption some customers experienced earlier this week.
reminder that aside from pulling off one of the greatest comebacks in AI model history - Google also owns about 17% of anthropic
that’s about $260 billion at market open.
there’s a crisis emerging that threatens 90% of neoclouds:
- the U.S. is banning chip exports and cloud rentals to china.
- which means there’s now excess supply flooding the market. asian neoclouds are pricing gpus at discounts.
why? china appears to have been a major source of revenue
- roughly 10% of nvidia rev was china. they reportedly sold $10+ billion to megaspeed alone.
now that’s being cut off and (so far) no one’s buying up the compute.
my expectation: this oversupply is temporary. if the compute is there either anthropic, openai or any other legit hyperscaler will buy it.
if it doesn’t get bought up then you know demand for AI isn’t growing and/or compute was overpriced.
I would expect B300 pricing to continue declining.
A staggering number of neoclouds have approached us @liquidcompute looking to remarket long-term B300 capacity originally committed to Chinese-affiliated offtakers into the US and Western Europe due to concerns of stricter legislation regarding compute servicing to China.
https://t.co/YPnqDRQ5xR
history is made twice today:
the worlds most anticipated video game and the worlds largest IPO
ironically both will do a fantastic job of removing people from their jobs.
happy Grand Theft Anthropic day!
the blowout from the Nvidia scandal will be felt across so many other company’s
microsoft makes $1.5 billion from selling compute to a *single chinese customer*
again this isn’t unlawful (yet) but it does shine a spotlight on a “grey market” that will get shut down if decembers bill is passed.
whether that’s “the right thing to do” can be debated but it’s pretty clear where this is headed
US gov wants nationalisation.
Microsoft $MSFT is nearing a deal to sell TikTok parent ByteDance an additional $500M in Azure AI services, primarily for access to OpenAI models.
The potential deal is creating tensions with OpenAI, which has raised concerns that Chinese companies could use its models to train competing AI through distillation.
Microsoft says it has found no evidence that ByteDance is doing so and believes the company primarily uses the models for software development.
ByteDance was already on track to spend roughly $1B annually on Microsoft's AI and cloud services as of June, excluding this potential deal.
While OpenAI restricts direct access for most Chinese companies, Microsoft's licensing rights allow it to sell OpenAI models through Azure to Chinese customers.
Source: The Information
man google is cooking the competition - 3rd major model in 20 days:
- Carbon 4 competes with fable and gpt 6.2. currently ranks 3rd (it’s close)
- Neon competes with Opus and terra
- Hydrogen competes with sonnet and luna (#1 best imo)
and that’s excluding the Flash, Decision model launches!
pretty clear they’ve rebuilt the entire pre-training / post training stack, dedicated more TPUs to training and tbh i’m so glad - google has all the data and distribution to win
all they needed were good, distinctive models and now they have them.
Gemini 4 Neon is rolling out now.
Sitting between Carbon and Hydrogen, it brings frontier intelligence and fast task completion for complex software engineering, knowledge work, and science.
Travis Kalanick just now on Jensen Huang scandal:
“i will turn up with Nvidia HQ with a silicon wafer-encrusted megaphone with nvidia employees and shareholders yelling to get jensen reinstated… bring jensen back”
ok so the latest nvidia situation:
- jensen wants to come back. board said no but behind the scenes there’s apparently a faction pushing to overthrow this.
- Q3 earnings was a huge success, ceo colette says “full steam ahead” but market’s not buying it - stock is down this morning.
- yesterday the SEC launched an investigation into nvidia misleading investors on china chip exports.
- the spy scandal *still hasn’t returned evidence* alice is a spy,
- the kill shot is said to have come from marco rubio
- neocloud financing will likely stop for smaller players, with bigger player contracts being honored.
- the next 12 months are already locked in via jensen, what happens after that is what the market is worried about.
imo new ceo will determine confidence in nvidia going forward. rene haas speaking later at tmrwX summit 👀
BREAKING: "A faction" of Nvidia, $NVDA, board members is pushing for Jensen Huang to return as CEO, despite the board publicly stating that it is not considering his return, per Semafor.
Jensen Huang has reportedly met with numerous tech executives and spoken with at least one Nvidia board member about the possibility of returning as CEO.