David
544 posts

David
@dd_tch
Off The Edge of The Earth and Into Forever, Forever. AI and Web3 positive
New York Joined Aralık 2009
121 Following428 Followers

@ellie_huxtable its awesome! waiting updates! god swear Atuin became for me first tool I install on my vps, servers, local machines, vm's etc
English

vibe coded personal task tracker with calendar, added mcp to it and voice calls with @retellai. It's really cool when agent actually calls you instead of silent notifications we never pay attention to. github.com/ddtch/lite-task check it out and grab if you want

English

@dabit3 So sick - one thing I’d recommend to try is their frontend design skill to avoid the blue/purple gradient that ai tends to go for in frontends
github.com/anthropics/cla…
English

Something I wanted to see if Claude Opus 4.5 could do: clone a fully functional Billion $ SAAS product and make it at least 100x cheaper.
The first product that came to mind was TypeForm because it's very popular, very expensive, and in theory, very simple.
The result is OpenForm: a polished + functional and Open Source Typeform clone at ~100x less cost, that can be setup and deployed in ~15 minutes. The agent building this ran for ~35 minutes.
Here are the details, technique, and the code:
English

top 15% of cursor users. I know that @AlexReibman should be definitely in top 10 or more. What about our Agentic star @braelyn_ai ???

English

@braelyn_ai That’s so true. Same I previously said about android phones it’s experience ruined by google
English

last 45 minutes I was trying to generate 5 sec videos by very specific and not so much scripts with Veo, @Kling_ai, @midjourney and @grok. Here are my thoughts:
English
David retweeted

This paper builds an agentic LLM that can run the whole data science workflow by itself.
It is an 8B model that plans work, reads structured files, writes and runs code, checks results, and iterates.
Standard “workflow agents” break here because fixed scripts do not adapt well to long, multi step jobs.
DeepAnalyze fixes that with 5 actions, Analyze, Understand, Code, Execute, and Answer, so the model can switch between thinking and doing.
Training happens in 2 stages, first single skills are strengthened, then multi skill reinforcement in live environments.
They also synthesize step by step trajectories so the model sees full examples of planning, coding, and using feedback.
Rewards are hybrid, simple checks like correctness and format plus an LLM judge that scores report usefulness and clarity.
The result is autonomous orchestration, choosing the next best action, and adaptive optimization, improving decisions from environment feedback.
Across many benchmarks, this 8B model beats most workflow agents and can produce analyst grade research from raw structured data.
---
Paper – arxiv. org/abs/2510.16872
Paper Title: "DeepAnalyze: Agentic Large Language Models for Autonomous Data Science"

English

@braelyn_ai Haha to fix a hook we need a hook that hook you up… few moment later half or react developers are hookers 😄 pardon my pun
English










