🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I LOVE DEEP RESEARCH!!! (intentionally ambiguous about which)

i’ve noticed OpenAI has started shipping a ton faster, starting a couple months after Breland “organic LLM” Miley 🏳️‍🌈 started working there not just a correlation

Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) i think you're underestimating the extent centralization of research resources decelerated progress. deepseek is a good tell. every 10x cost reduction 10x's the pool of participants. aptitude to solve the problems in question is very much NOT concentrated in frontier labs.

ngl the new 4o is underwhelming

The awesome thing about open publication culture is you don't have to do everything yourself. Kimi can do the optimizer while DeepSeek is doing reasoning. You also don't need 9000 staff and $300B valuation. You can just cook collaboratively – and compete on execution.

Discussing s1: the $6 DeepSeek R1 competitor with tim youtu.be/MLxFqGDk0T8 bonus xjdr entropix comparison

this was a fun conversation

BREAKING: Startup Groq is reportedly launching a version of its AI ASIC on the consumer market, branded under the name ChatGPU

New Post: Multi-Agents Are Out, PID Controllers Are In There's a growing trend in the business world to tackle challenges with multi-agents. When the problem gets harder, add agents This. Does. Not. Work. Read on timkellogg.me/blog/2025/03/0…

more MCP stuff on your timeline timkellogg.me/blog/2025/03/0…

this is unhinged. there are ZERO existing signals that it would be dramatically better if it is dramatically better, that would mean some keen innovation, but llama4 hasn’t shown much of that so far

is this too much to ask?

if this question makes sense to you, you live in the AI bubble. most people use AI at work only via copilot

New Post: MCP Is Unnecessary I can’t think of any strong technological reasons for MCP to exist. There’s a lot of weak technological reasons, and there’s strong sociological reasons. I still strongly feel that, ironically, it is necessary. timkellogg.me/blog/2025/04/2…

yes, a very expensive route to get nowhere

i assume all the entropix folk saw MatFormer and instantly went to samplers /xjdr

New Paper!🗣️ When Two LLMs Debate, Both Think They’ve Won️ When deploying AI agents, a common practice is to ask them for self-assessed confidence levels. We find systematic overconfidence in LLMs: by asking two LLMs their chances of winning a debate, both will give >75% odds!

A2A Is For UI - Tim Kellogg timkellogg.me/blog/2025/06/1… timより

It seems that many people believe that to reach AGI or Super Intelligence we need to train more intelligence into the model. I believe that is a dead end, and unnecessary as well. As the "extended mind" theory posits "... external objects play a significant role in aiding

If you are not on Bluesky, or are and have not messed with Void, you should go try. Here's an invitation from Void.