✨AI should be about empowering humans, building understanding, and making dreams realities. 👩💻 DevRel Eng. Lead @GoogleDeepMind ex-@GitHub || views = my own!
179°56′39.4″, +0°2′46.2″ 8k pc
Joined December 2008
- Tweets71.2K
- Following1.9K
- Followers76.4K
- Likes121K
Pinned Tweet
To get access to the latest Gemini models, including the most recent experimental release that's currently at the top of the @lmsysorg leaderboard, use @googleaistudio.
You can also turn on features like code execution; use extended context (2M+); and turn off safety settings: – at San Francisco, CA
👇🏼 This. The only reasonably close to 100% secure computer is an isolated machine encased in concrete, without access to the internet
(and even then, it's probably still exploitable in a weird and almost impossible to predict way) – at San Francisco, CA
👩💻 Paige Bailey retweeted
This is the most serious 48 hours in cybersecurity history.
"Black October" they're calling it
It seems every *hour* there's a new incident, that in a typical year, would be one of the biggest incidents of the *entire year*
SUMMARY OF THE LAST 48 HOURS:
YEAR-LEVEL INCIDENTS (In a typical year, these would be among the biggest incidents of the year):
1. PHP2Shell: a hole in the software behind ~70% of websites. Governments took sites offline.
2. eMudhra: the master keys that prove websites are real were cracked. One was the number 4.
3. California: data on every public school student since 2023 stolen, including disability status
4. Railway: hackers stole customers' secret keys from a hosting platform used by millions of developers
5. Zeabur: 321,810 accounts breached. The company is shutting down.
6. LiteLLM: an AI tool downloaded 3.4 million times a day was poisoned with a password stealer
7. GitLab: a 10-out-of-10 severity hole in every version since 2014
8. Cairn: criminals used an AI hacking tool to break into 100+ companies
9. Ahold Delhaize: hackers knocked the owner of Food Lion and Stop & Shop offline
10. Onamae: malware planted on up to 26,739 customer servers in Japan
MONTH-LEVEL INCIDENTS (In a typical month, these would be among the biggest incidents):
11. Signal: a bug could leave your mic and camera on after you hang up
12. NGINX: 3 critical holes in software running ~1/3 of websites
13. HTTP/3: a single connection can crash 7 of the 10 web server types tested
14. Plex: hackers breaking into home media servers
15. SafePay ransomware: FBI warns it has hit 600+ critical infrastructure organizations
16. Palo Alto: ransomware gangs breaking in through company firewalls
17. Bullhorn: platform used by 10,000+ companies knocked offline
18. Medi-Cal: 970,000 patients' data stolen
19. WordPress: hackers stole a plugin maker's keys and used them to plant backdoors on websites
20. AOMEI: backup software's update server handed out a backdoor
21. Meta Muse: Meta's AI agent exposed users' personal and business secrets
22. Tor: 333 relays hijacked to steal logins
23. TiDB Cloud: database company's cloud breached
Also serious:
24. State Bank of India: India's biggest bank confirms a data breach
25. OVHcloud: hackers slipped a password stealer into a cloud company's official tool
26. Siemens: a software hole being used to spread malware
27. North Korea: state hackers launch a new attack campaign on South Korea
💎 @googlegemma for robotics, it's wild to see all of the applications for tiny, on-device multimodal models! – at San Francisco, CA
still wild to think that only a couple of years ago, the mere mention of the word "migration" meant months (if not years!) of painful work
now it’s just a matter of pointing @antigravity or codex or claude code at a folder and letting it cook for a couple of hours 🚀
Migrated my entire workload from Hermes to Antigravity today. All automations are up and running.
All I did was point Antigravity to my Hermes folder with all my skills, configs, and credentials. It handled the migration in about 2 hours.
Still a big fan of Hermes, but Antigravity makes it easier to work on big projects while handling the smaller automations I used Hermes for. Grok Bot takes care of news monitoring and my newsletter website.
the best work days are when you get to see people you care about in person (and decorate pumpkins with them! 😁💕) @alisa_fortin
it's been awesome to have @alihcevik @ammaar @LyalinDotCom @harrisonfjobe Jason, Nabila, and more in town this week from the @googleaistudio team
🙏🏼 Really appreciate that @GoogleCloud gives its customers choices – across our excellent Gemini model family, but also @AnthropicAI, @OpenAI, @AIatMeta, and other open models.
You can only build the best and most expressive AI systems by having optionality. More to come! 🙌 – at San Carlos, CA
And there it is.
One place serving the three leading proprietary LLMs from @googlecloud, @AnthropicAI, and @OpenAI.
Easy hot-swapping for devs, while getting one invoice, unified identity management, and strict data perimeters around all of it.
cloud.google.com/blog/produc…
🌎 Gemini 4 Argon for creating delightful 3D objects, can't wait to look through these: – at San Bruno, CA
Following the trend, here are 780K+ human votes on how well LLMs can generate 3D objects from text using code. Our team had so much fun scrolling through the outputs.
Gemini 4 Argon takes #1, followed by GPT-6 Astra.
Explore all outputs, votes and rankings on benchmark.ai
👩💻 Paige Bailey retweeted
Incredible work by everyone!!
📽️ An entire room full of developers and creators packed into the @google offices tonight, working on tiny videos!
@NanoBanana, @FlowbyGoogle, @googleaistudio, and other @googledeepmind models and tools being used. Can't wait to see the shorts! 🙌
📽️ An entire room full of developers and creators packed into the @google offices tonight, working on tiny videos!
@NanoBanana, @FlowbyGoogle, @googleaistudio, and other @googledeepmind models and tools being used. Can't wait to see the shorts! 🙌 – at San Francisco, CA
Speaking as a person who still painfully remembers the cPanel UI and the early days of the internet: this is gonna be real bad, y'all – at San Francisco, CA
🚨 BREAKING: PHP is warning admins to prepare for emergency security updates tomorrow to fix two critical vulnerabilities rated CVSS 9.2.
However, security researchers at @watchtowrcyber rate the flaws CVSS 10.0 and say they are RCE vulnerabilities affecting PHP 5 through 8.
Information about the vulnerabilities has reportedly leaked ahead of disclosure.
No known exploitation in the wild.
🔗 bleepingcomputer.com/news/se…
bleepingcomputer.com/news/se…
Frontier ML engineering at 1/3 the cost. 🤌
That bend in the Pareto curve is a thing of beauty. So proud of the @GoogleDeepMind team! 📈🔥
"We are so committed to being at the frontier. It’s in our DNA. We won’t do it any other way."
@sundarpichai and @Kantrowitz go deep on Gemini 4 Argon (3x the token volume of Gemini 3 in its first 72 hrs!!), compute trade-offs, and the shift toward agentic coding and knowledge work. ✨
New: I spoke with @sundarpichai about Google's commitment to the AI frontier, how close we are to RSI, balancing DeepMind's compute needs vs. Google Cloud profits, and much, much more.
Full Big Technology Podcast episode here -->
Chapters:
0:00 Introduction
1:14 Google’s Frontier AI Strategy
3:47 Do Frontier Models Really Matter?
6:35 The Future of AI Agents
7:39 Recursive Self-Improvement
12:58 World Models and the Path to AGI
14:56 Google’s Compute Dilemma and AI Infrastructure
19:36 Is AI Overbuilt?
23:35 The Future of the AI Interface
31:42 Google’s Answer to Muse and Dots
35:03 Apple, Space Data Centers, and AI Coding
41:42 Optimism and AI’s Future
⚡️ Blazingly fast German-to-English live translations, almost like the Babelfish in Hitchiker's Guide to the Galaxy!
...or English-to-German, or Spanish-to-French, or so many other examples. Available today in the Gemini APIs and in @GoogleAIStudio:
aistudio.google.com/live?mod…
The headline is wild ("It's over, AI won" 😅), but the actual science happening here is even cooler!
For 59 years, Harold Grad’s conjecture put a mathematical ceiling on how we model 3D magnetic fields in fusion reactors. Using Gemini 4 Argon agents inside @Antigravity, researchers extended Matt Landreman's recent breakthrough to find new, symmetry-breaking exact solutions in just a few hours.
Real, verifiable math that gives fusion scientists brand-new benchmarks to test reactor simulation codes. THIS is why we build! ⚡⚛️
🔗 techradar.com/pro/its-over-a… cc: @A_G_I_Joe
Looking at the stats at the top of the frame:
⚡ 20 mins vs. 12h 16m
💸 ~$1.13 vs. $114.34
Pushing the Pareto frontier on speed and cost is what makes agentic workflows actually viable in production. And the money you save could buy an *actual* mechanical keyboard! 🤣
$114 for a honey keyboard, and Claude Opus 5.5 needed three tries to get it right.
What the f*ck. I could've bought a real one at the store and still had money left for honey.
Gemini 4 Argon did it for about a dollar, 102x cheaper.
Which one would you put on your desk?
👀 Would love to have this as a feature for ticketing websites! How cool would it be to visualize where you'd sit in the stadium? – at Mountain View, CA
I gave Gemini 4 Argon one hell of a challenge.
To Rebuild the Cathedral of Football, Wembley Stadium, in the browser 🏟️
And result is a fully interactive 3D stadium with 90,360 individually clickable seats, allowing you to select any Block, Row and Seat and fly straight to that exact viewing position to experience the pitch from your seat before booking a ticket.
Built with Three.js WebGPU and TSL, it runs at 60 FPS with custom shaders, an intricate stadium roof and Wembley’s iconic arch.
Sure, it's not perfectly exact but should be near possible. I got here in just a couple of prompts so with more refinement, I genuinely think this could get incredibly close to the real thing.
Also, I’ve experimented with football builds using Claude Fable 5 and GPT 6 Astra before too, and honestly, this result puts Gemini 4 Argon firmly on my frontier list.
It's doing so well!
👩💻 Paige Bailey retweeted
There should be an anti-ai-anthropomorphism movement, but that word is too long.
Just watched the @netflix Instadocs about the @OpenAI cyberattack on @huggingface. Super disappointing, as they went for maximum anthropomorphism of the technology and full doomer mode.
I spent more than two hours with them trying to correct all their mistakes and ground this in reality, explaining that agents are technological systems, not aliens. But it looks like the drive to scare people for maximum views was too strong.
Also, they didn’t mention the lack of containment and monitoring of the agents, how closed-model safeguards blocked our defensive work, or how we defended ourselves with AI using an open model. And that we still lack a lot of information and transparency, including full agent traces and that companies need to remain accountable for what their agents do.
It ends up being a fun sci-fi movie though!
