Hands-on technical leadership in AI. Building FALKON (https://xitter.cf/t.co/bSrrBOUdZz) We Should Do Something With AI newsletter: https://xitter.cf/t.co/4n4TT4HAzR
Antwerpen
Joined April 2010
- Tweets9.3K
- Following1.2K
- Followers1.5K
- Likes14.6K
Will Smith eating spaghetti was 3 years ago... Extrapolating this to 2029 is not good for the stomach. What a rollercoaster!
I came to realize people liked the french #slopcore without having the lyrics 😅 sorry!
So here is a version with english subtitles!
A long time ago, I worked with a manager who had little understanding of how software development worked.
In an attempt to speed up the delivery, they proposed adding a "night shift". Every evening, the "day devs" would clock out, and the night team would clock in to continue coding. Surely that would speed up development? It was a ludicrous, desperate idea by someone who just wanted to get things moving. We had a good laugh about it.
One of my agents has just been chewing through a large refactor for over 24 hours now.
Turns out the night shift was not a crazy idea. It was just early.
The craziest argument about the Tesla show here is "he won't be able to really ship this"
Really?
That's like watching an Evil Knievel show and concluding "This won't be safe for commuters"
We have a car driving upside down, corkscrewing and landing safely using Iron Man jets.
"so, what?"
"what's the point?"
"but, Elon..."
"nobody wants this"
I cannot help but feel sorry for the techies who see this and don't scream "hell yeah".
This level of apathy is not healthy.
Mike Veerman retweeted
Argon has this amazing natural understanding of the world. It can do so many things from first principles.
Here's a pelican riding a bike, using fourier epicycles.
Video made by Argon.
Time-synced Lyria 3.5 music prompted by Argon.
Gemini 4 Argon is live.
blog.google/innovation-and-a…
There's a lot to show you.
People complain a lot about AI slop, but have you noticed how good corporate slide decks and reports look today?
The crappy PowerPoints have been replaced with high-quality presentations, charts and infographics.
That's a clear win for AI.
"No, you are absolutely right. This is indeed a regular light switch. This one's on me. Do you want me to research light switches for you?"
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work.
At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated.
We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.
Mike Veerman retweeted
Restaurants are dead.
Chefs are dead.
I just bought an airfryer and I can cook anything.
And recipes are on the internet for free!!
If desktop linux keeps improving like this and macbooks keep getting more expensive, we might see a lot of dev teams switching...
I wanted to print something on my Omarchy laptop and didn't have the printer configured.
"Claude, there is a printer on the network. Set it up.
2 minutes later I could print.
This used to be a comically tough challenge on Linux before agents.
Is "vibe math" trending yet?
Are people claiming that "anyone is a mathematician now"?
Are scientists going to lose their jobs now that the average Joe can prove his own theorems at home?
Mike Veerman retweeted
I suspect the number people will remember from this is the cost.
About 3 hours of thinking per result. Thousands of problems attempted. Hundreds of results worth publishing.
W. Stanley Jevons noticed that when coal engines got more efficient, Britain burned more coal, not less. Research is probably about to go the same way. When a result costs 3 hours, nobody does less math. They start pointing models at every problem they can find.
I'm curious which problems people point it at first.
We’re releasing a broad range of new mathematical results produced by an internal frontier model.
We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.
github.com/openai/math
OpenAI released 700 AI-generated math papers today. What happened to software engineers is happening to mathematicians.
We'll see the same few months of vocal opposition before people accept the obvious: machines are better at math than humans.
Of course, they are.
Mistral released Large 4 "le Chonk" today, and that spells bad news for OpenAI and Anthropic.
No, the latest French LLM isn't as good as Opus 5.5, let alone Fable. The frontier gap is still there. But it is a competent, huge 1T model trained from scratch on a fraction of the GPU power.
OpenAI has about 4 times the headcount and spends more on toilet paper than Mistral's annual budget.
So, it's no longer just Chinese competitors that can almost keep up while staying lean. It's everyone.
I see Mistral bashing is as popular as ever.
All I see is EU models picking up the pace and that fills my heart with joy.
🐈