onel.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

onel.bsky.social

Founder askmanu, helping developers get answers

14 posts 16 followers 74 following

Posts 10 Comments 8

The pricing for the new OpenAI GPT 4.5 is crazy. They said it's a big model but holly. We haven't seen this prices since 2021

submitted 8 days ago • 0 comments

The new Anthropic Sonnet 3.7 is already available on askmanu.co. The model tops the coding benchmarks and together with the most up to date docs from askmanu, you get he most powerful code assistant. 👇

submitted 11 days ago • 0 comments

New research measuring exactly how much LLMs loose performance with longer context. I wish we could just dump everything in context, but it just doesn't work. Link: arxiv.org/abs/2502.05167

submitted 22 days ago • 0 comments

Amazing keynote from Michael Jordan (the statistician not the NBA player) about the role of AI. We can start adding value now to people / companies and not wait for the "super intelligence on the hill".

submitted 23 days ago • 1 comment

He is a man of focus, commitment and sheer f***ing will www.youtube.com/watch?v=T_2f...

submitted 38 days ago • 0 comments

JPMORGAN’s desk calls #DeepSeek “the first major fundamental challenge to Mag7 and the AI theme .. “.. it is too early to abandon Mag7/AI theme especially with earnings this week. That said, I do think this accelerates the timing on when the market will want to see an ROI on AI investments ..”

submitted 40 days ago • 4 comments

Crazy if R1 is the reason for this

submitted 40 days ago • 0 comments

For coding tasks, using 𝗗𝗲𝗲𝗽𝗦𝗲𝗲𝗸 𝗥𝟭 𝗳𝗼𝗿 𝗿𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 and then 𝗖𝗹𝗮𝘂𝗱𝗲 𝗦𝗼𝗻𝗻𝗲𝘁 𝗳𝗼𝗿 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 apparently gives the best results. aider.chat/2025/01/24/r...

submitted 40 days ago • 1 comment

This deep dive on Netflix purposely turning movies into barely distinguishable forgettable slop is so insightful, researched & brutally hilarious. A must read if you want to really know what’s happening to the business AND culture of movies.

submitted 70 days ago • 74 comments

Wrote up some notes on Microsoft's new Phi-4 LLM. They trained it on a LOT of synthetic data, and the details of how and why they did that are really interesting. simonwillison.net/2024/Dec/15/...

submitted 83 days ago • 4 comments