mattwallace
@mattwallace
Followers
651
Following
1K
Media
716
Statuses
7K
techie, dad, author, inventor, perpetually curious; CTO & Builder all-in on #AI https://t.co/0MDUTxXXzb
The Arena
Joined January 2008
cat https://t.co/klLNTJOm1H | cllm translate to french > docs_french.md https://t.co/Taz4ulLFNB what's next?
github.com
(C)ommand-line (LLM) calls. Contribute to m9e/cllm development by creating an account on GitHub.
1
0
6
Four years ago, an SVG of a pelican riding a bicycle was not on my list of things that would make me go, "what a time to be alive," but here we are.
1
0
0
the other hall of fame, but not an SML. Pricelessly bad.
omg, I love GPT5 but someone on my team pasted this and I am *dying*, please send oxygen.
1
0
0
Man, this gets better every time I read it.
😂😂😂 The next time someone tells you SMLs are the future of agents, this gem from one of my engineers working on our evals system.
0
0
0
Hey @demishassabis my friend Brian Weissman (of Magic: The Gathering fame) and one of the cofounders of GGG / @pathofexile thinks an "AlphaMagic" cannot be built which can beat a pro-tier human player at old school magic. Want to prove him wrong?
0
0
0
At a certain GPU spot, this is gold. I created DF11 compressed weights (lossless ~32% weight reduction) of the new @Alibaba_Qwen Qwen-Image/Qwen-Image-Edit: https://t.co/zoNjqH2hA4 /
huggingface.co
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
0
0
0
Did a DFloat11 compression of Qwen-Image-Edit-2511 https://t.co/rO2I7FEVqe
@Alibaba_Qwen ; thanks Qwen team & the DFloat11 authors.
huggingface.co
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
0
0
0
9 years of flight, but also, at avg pluto distance of ~3.7B miles, this is 5.5 h just of transmission time at the speed of light.
The math on this image is insane. New Horizons transmitted at 2,000 bits per second from 3 billion miles away. Slower than a 1990s dial-up modem. It took 16 months to download all the flyby data. The spacecraft had to hit a target box 100km wide, arriving within 150 seconds
0
0
0
💯 This is so good.
High school student uses AI to discover 1M+ objects humans missed in astronomical data. Head of NASA openly recruiting him through Twitter with a fighter jet ride included. All my worlds colliding. I love everything about this.
0
0
0
If you're wondering whether saturating ARC-AGI-1 or 2 means we have AGI now... I refer you to what I said when we launched ARC-AGI-2 last year (which is also the same thing I said when we announced ARC-AGI-2 was coming, in Spring 2022, before the rise of LLM chatbots)... The
96
99
1K
Tension based scale attached to luggage carts. Make it happen world
0
0
0
🙌 💯 Super grateful for the Olmo team and this effort. We should all shout this from the rooftops. Open weights are awesome and useful and help, but *this* is the "open" that maximizes progress for humanity.
Olmo 3 is one of the most valuable open research artifacts to ever be released. Although Olmo 3 models are slightly behind state-of-the-art, their value goes beyond the models themselves. The artifacts for Olmo 3 give anyone the ability to conduct rigorous experiments with
0
0
1
tldr: 2 guys with a laptop removed ~70% of the world’s compute bill for free, cuz why not
You can now train LLMs 3× faster with no accuracy loss, via our new RoPE and MLP kernels. Our Triton kernels plus smart auto packing delivers ~3× faster training & 30% less VRAM vs optimized FA3 setups. Train Qwen3-4B 3x faster on just 3.9GB VRAM. Blog: https://t.co/j3HpsmzrCw
37
234
5K
Yann LeCun says LLMs are not a bubble in value or investment; they will power many useful apps and justify big infra The bubble is believing LLMs alone will reach human-level intelligence Progress needs breakthroughs, not just more data/compute "we're missing something big"
172
310
3K
Tim compressing the old adage appropriately
@minhsmind Everyone overestimates what will happen in the next 3 months and wildly underestimate what will happen in the next 3 years.
0
0
0
@QuinnyPig I mean I'm just a dude who got curious here and I could have gone off the rails but it really looks like AWS is selling like way over $1B in anthropic inference in 2025 and... I'm just thinking I must be doing it wrong. Is it possible no one noticed this at that scale?
0
0
1
This was 90 minutes of my life I did not have to sacrifice to this, but it's been interesting. @QuinnyPig You think people should get refunds for bedrock usage where the default prompt gaslights the model? 🤯
1
0
1