PREDICTION 2026-2027 will bring a new era for opensource AI An era that will be DOMINATED by American opensource labs pushing the frontier of open models >The gap between close & open models will get narrower, not widen as many speculate This tweet is for history, bookmark it
@theahmadosman
-
Reviewer feedback on Policy A and options for acceptance
By
–
This is the GLM-5V which is different from GLM-5 btw
-
AI Research Paper and ICML Conference Update
By
–
People who got tricked into buying a Mac mini by the Mac grifters now have a way to actually use them for local models
-
Academic Reviewing Policy for Conference Submissions
By
–
Yeah, you should DM @willccbb your latest research repo and lock-in the position before someone else snipes it
-
ICML Conference Policy Clarifies LLM Use Rules, Not an LLM Ban
By
–
No, this isn't the future of heterogenous hardware (e.g. combining NVIDIA GPU w/ Mac Studio Unified Memory) It's TOO SLOW for that This would only work if the entire model is offloaded to the GPU, otherwise it becomes a massive bottleneck
-
ACM Turing Award Announcement Anticipation
By
–
Just wrote a post about it Some will say this is good for mixing and matching hardware, ignore them, as this will be way slower than Unified Memory Besides that, great option for single GPU inference
-
User requests Gemini to create and populate Google spreadsheets
By
–
This will probably be great for Large single GPUs (e.g. RTX PRO 6000) You’re limited to 40Gpbs initially (during model loading) but then once the model is fully loaded on the GPU it should be extremely faster than Unified Memory speeds for inference
-
Anthropic Partners with SpaceX for Increased Compute and API Limits
By
–
let’s gooo gonna do F# and Haskell next
-
South Korea’s First Robo-Monk Arrives, Impresses in Fasting Test
By
–
rewriting Claude Code source code in COBOL through a locally running Hermes Agent with MiniMax M2.7 pic.twitter.com/rWAKB06lid
— Ahmad (@TheAhmadOsman) 1 avril 2026rewriting Claude Code source code in COBOL through a locally running Hermes Agent with MiniMax M2.7
-
AI News: OpenAI Agents, Anthropic for Finance, Mini Data Centers
By
–
Glad so many people are finally Buying a GPU