⚡+👨🏾💻= 🤖 @GoogleDeepmind | Prev: Gemini for @Android, Health @GoogleDeepmind, @Facebook, Founder, @Google @GoogleMaps @Apple, @SCSatCMU, Vermont, 🏀
California
Joined August 2008
- Tweets2.7K
- Following751
- Followers2K
- Likes3.9K
🥇 Gemini is #1 on my favorite coding benchmark for the first time!
More to come soon :) 🏎️
Gemini 4 Carbon is now available in Cursor!
At high effort, it's the top model on CursorBench at 60.4%
4
226
One of the most surprising experiences I had earlier in my career with Larry was observing how often he would make a comment/prediction that sounded completely crazy, including to experts in the domain, only to be proven correct weeks, months or years later.
So many of Google's most game changing products were a direct result of his product/market intuition & willingness to invest for the long term.
Even the modern day AI Chatbot interaction was something he stated as the vision for Google more than 20 years ago.
Larry Page reportedly has a number for how much computing will end up in space: five nines. As in 99.999%.
@patrick_oshag heard it secondhand: "Someone asked Larry Page how much of compute will be orbital, and he said five-nines."
@martin_casado's reason it might happen: "It's just basically regulatory." He thinks there are 500 proposed bills to stop data center building in the US, Texas included.
From @InvestLikeBest.
237
People often forget that power, compute, & systems are *the game* right now in AI.
We're going to see a lot of step change innovation on efficiency beyond the model layer.
If you're not in control of the key pieces, it's going to be a difficult next 6-12 months.
1
2
166
The most inspiring thing about Senator Elect Abdul El-Sayed’s campaign is that he didn’t shy away from who he was, his background, or his beliefs.
He’s the latest in a series of political winners who are genuinely authentic leaders. You can disagree or agree with them, but you know what they stand for.
So proud to be American & excited for the next-era of American politics.
Trying to read through the California Voter Guide with my daughter.
How does any normal voter understand this?
There are literally multiple propositions designed to cancel each other out.
Good opportunity for government to leverage AI to create a more informed voter base!
139
Nice! Good new independent benchmark.
Gemini Carbon at #2!
Today we’re releasing GDPx, a benchmark for measuring agentic capabilities across the most valuable domains.
GDPx is a true frontier-level benchmark: even the best model scores just 12% (1/7)🧵
1
11
445
This is not quite right.
The opportunity for local inference is not for everyone to get a new computer to run AI, it's for them to run inference on the devices they already own.
Everyone is going to have local compute/memory on their existing phones, tablets, PCs, etc...
Maximizing the utilization of that existing compute for AI inference is additive to global compute supply & likely necessary for the industry to keep up.
Tesla is a great proof point of this with FSD. Train a big model in the cloud & distill it to run on local compute in a car. What they've been able to get out of 8 year old compute hardware has been incredible.
Future is bright for local AI compute!
sam: local inference is a terrible allocation of scarce resources
“a lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we're going to be able to run them on our computers, and so we're not going to need all the stuff in the data centre
but from a global allocation perspective, actually it's quite terrible to have everybody, like, have a bunch of the world's limited supply of DRAM sitting in their own computer and being used five percent of the time
you want maximum utilisation in the cloud if you're trying to deal with this very supply-constrained environment
somehow many people's intuition is the opposite of that”
142
Best part of this is the speed of iteration! Huge leap in just 2 weeks from our Argon release.
Shows how much promise there is in the Gemini 4 model series.
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
107
Haters are claiming this demo will never have practical application.
They're missing the important role of a leader to inspire his team + world.
Demos + events like this combined with a great product is why no one can compete with Tesla.
83
Quantum will be multi-trillion in the ***next few years***??
Love Zohran but who is advising him on these investments :)
Quantum is poised to become a multi-trillion dollar industry in the next few years. We think New Yorkers should share in that profit.
That's why we negotiated that New Yorkers get board seats and a 15% stake of intellectual property that comes out of NYU's Quantum Fabrication facility.
Because all New Yorkers should benefit from the economic success of this city.
151
Awesome progress!
Can confirm Gemini 4 Argon is big leap in coding - great to see this being validated by the industry's best coding benchmark.
Give it try in Cursor!
Gemini 4 Argon is now available in Cursor!
It scores 52.8% on CursorBench at default effort, on par with Opus 5.5.
Priced at $2/M input and $10/M output tokens.
1
3
243
I think the smarter play for Starlink Mobile would be to enable mobile phone companies to go directly to consumer.
Be the network for Google Fi, Apple Wireless, Samsung, etc... to provide cell service directly to their users.
We announced an agreement to acquire a nationwide low-band spectrum license portfolio that will pave the way for @Starlink to become a major mobile carrier in the US.
With this new spectrum and our Gen2 constellation, Starlink Mobile can ensure Americans have access to high-speed mobile broadband no matter where they are → spacex.com/updates#starlink-…
117
Two of the most innovative companies in the world coming together for a launch 🚀 🛰️
I remember the first meeting where the team pitched the idea of putting compute in space - exciting idea!
Today's successful launch with our partner @planet prototype satellite shows how far we've come...and how far there is to go. Thanks for the lift, @SpaceX, another spectacular booster landing too🚀
112
Back at the Frontier!
It's been amazing to use this internally, I was able to make the largest code changes of my PM career thanks to Argon - bringing ideas to life in days that were estimated to take weeks.
Can't wait to see industry feedback.
Lots of discussion out there about our next model(!), so I wanted to give an early look as soon as possible. Introducing Gemini 4 Argon!
It shows frontier performance in complex workflows, cyber defense and software engineering. Teams are using it extensively at Google, from coding to quantum computing, great feedback.
Here’s a look at the benchmarks:
1
160