←
Tom Cohen ✨1,848 posts
Tom Cohen ✨ avatar

Tom Cohen ✨

@tomcohen

10% memes, 90% weapons grade autism. Wrangler of model intelligence @ Labyrinthine.

◷ Joined March 2022▧ 296 Photos and videos
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·15h

It’s sad to see people who diss the gpt 6.2 generation instead of embracing the weird and beautiful mind that it is. A really unique model and has real “sparks of AGI” moments.

0:12
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·15h

As with all things, graphics generation goes into the transformer hole

I almost forgot my favourite GPT-6.2-Astra generation - this is SVG (yes SVG!!) of a peacock - no other model came close to the precision, lifelike nature and detail of this generation, incl. Fable

0:55
MahsaAlert avatar
⟳ Tom Cohen ✨ retweeted
MahsaAlert@mahsaalert·Nov 17

اگر همین الان اینترنت قطع بشه چی؟ ۱. نقشه آفلاین رو از قبل دانلود کن ۲. شماره‌های ضروری رو روی کاغذ بنویس و با عزیزانت محل ملاقات بذار ۳. باتری رو مدیریت کن و خبر تأییدنشده رو بازنشر نکن آمادگی، قبل از بحران شروع می‌شه. #مهساالرت #قطع_اینترنت #mahsaalert

1:01
Avichal Garg - Electric ϟ Capital avatar
⟳ Tom Cohen ✨ retweeted
Avichal Garg - Electric ϟ Capital@avichal·Nov 16

I can't believe the banks are down and I can't access my money. Can someone please invent a global, 24/7 network, with a redundant ledger that enables self-custody of my funds and is resistant to state-level attacks?

#BREAKING: Chase, Bank of America, Wells Fargo and U.S. Bank hit by nationwide outage, users report.

Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 17

Absolutely hilarious to see people in the comments take this at face value and thus as a bearish sign. He’s the CEO of one of the largest buyers of chips, he needs leverage in negotiations. He signals to vendors “we have alternatives” because it’s a good move. 🤷🏻‍♂️

Sam Altman says chipmakers are double counting OpenAI demand, each assuming OpenAI will buy all of their chips “Sure. For a long time, there was basically one chip provider in the industry. And so when people were counting wafers, they could ask companies like ours what demand was going to be. And they could, I think, get a sense of where end user demand for the chips they could make would be.” “There are now many people making chips. And many of those people think they’re going to sell most of the volume to us and Anthropic and others. And we have some limited amount of demand and data centers and everything else. And I think there are many people making chips all talking to the silicon supply chain assuming that we’re going to buy all of their chips.” “And I think the demand, the data center build out, that stuff is solid. But we’re not all gonna be able to or need to buy every… all of our supply from every chip vendor.”

1:01
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 16

Weird how Anthropic started cracking down on distillation and this release ends up so far behind.

📢 Meet Qwen4-Max, your new AI teammate for work, creativity, and discovery! Qwen4-Max is built on a new MoE architecture with native vision and a 1M-token context window. Highlights: - Outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over 60% in our internal evaluation. - Advances real-world productivity through greater proficiency with professional tools, stronger long-horizon execution, and higher-quality deliverables. - Handles real tasks, including 3D modeling, game development, manufacturing, finance, law, healthcare, and frontier scientific discovery. - $1.5/M tokens for input, $4.5/M tokens for output, and $0.075/M tokens for implicit caching — all lower than Qwen3.8-Max, with implicit caching down by 70%. The API is live on QwenCloud. You can also try Qwen4-Max on Qwen Studio, QwenWork and Qoder. Start building today! 🚀 - Blog: qwen.ai/blog?id=qwen4-max - API: qwencloud.com/models/qwen4-m… - Token Plan: qwencloud.com/pricing/token-… More to come! The open weights of smaller Qwen4 models will be released. Stay tuned. 👀

Post attachment
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 14

The fact that he mixed up the axes here makes this art.

Over-simplistic and brainwashing. Fixed the Axes "Boldness" is just what rich kids call gambling when their dad owns the casino. Since 1971, the economy hasn't rewarded brains or innovation - it rewards owning assets you bought with someone else's head start. Keep hustling, though. The "bitter professors" are busy writing the code the "empire builders" will take credit for.

Post attachment
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 12

Amnesty really have decided to go all in on supporting terrorists and gangs, huh? Weird timeline

The Prime Minister of Haiti, Alix Didier Fils-Aimé, and the President of El Salvador, Nayib Bukele, must ensure that all operations conducted by Salvadoran forces in Haiti comply with international human rights law. bit.ly/3UOTZLU

Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 12

Wild stuff. Really nice work.

🚀 Qwen, Kimi, GLM are all betting on linear attention. But it's still slow: the recurrent state is read & rewritten every token. Meet LeapQuant: the first training-free method to make 8-bit or 4-bit recurrent states work. 📈 Across Qwen, Kimi, and GLM models: • Matches FP32 accuracy • 4.25× faster linear attention kernels • Up to 2× higher inference throughput • Cuts memory by up to 56%

0:34
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 11

Perfect

Finally, AI writing often has no prioritization of the points it wants to make. This results in sentences that are devoid of meaning or sentences that have multiple distracting points. 4/5

Post attachment
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 10

"This tech existed for 3-4 years now and didn't cure cancer yet, can we all agree it's useless?"

So we've had LLMs for a few years now, right? Where's the big thing they've done? Where's the cancer cure? Where's the unlimited abundance? Honestly, can someone show me one great thing LLMs have done for society?

Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 9

They’re just showing off at this point

We wanted to do more for the people who’ve been with us the longest. So with Fable 5.2, the longer you’ve been subscribed, the more Fable resets you get.

Post attachment
Tom Cohen ✨ avatar
Tom Cohen ✨@tomcohen·Nov 9

The fact they’re calling it 5.2 and not 5.5… I’m scared.

Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges. They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.

0:19