Florent Daudens

Building to bridge creators and AI
AI & Media
ex @huggingface @radiocanadainfo @ledevoir

[Florent Daudens @fdaudens]

Kimi K3 beating Claude Fable 5 and GPT-5.6 Sol on several benchmarks shows that the cadence has fundamentally changed.

The rule of thumb was that open models lagged the closed frontier by several months; it now looks more like weeks.

[Florent Daudens @fdaudens]

Do you still use skills frameworks like Superpowers for complex tasks with GPT-5.6 or Fable?

I’m wondering whether this kind of scaffolding can sometimes get in the way of the latest models. Curious what’s working for you.

[Florent Daudens @fdaudens]

So who wins between GPT-5.6 Sol and Fable? After a few weeks of early access, GPT-5.6 is my daily driver.

Fable feels more adventurous, focused and weirdly narrower.

Sol feels more broadly useful and circumspect: closer to the brief, more reliable.

Demo https://t.co/6so5A4AJv8

[Florent Daudens @fdaudens]

GPT Live is seriously good. GPT Live with interactive widgets is next level.

After a few weeks with OpenAI’s new voice model, I turned voice on by default and there's no going back.

[OpenRouter @OpenRouter]

Introducing the OpenRouter MCP, live model intelligence right inside your agent

Your agent builds and ships, but when it comes to choosing the right model for the right job, it guesses from 6 month old training data

Watch it pick, price, and test the right model:

[Florent Daudens @fdaudens]

YouTube is officially in World Cup mode with a tiny soccer-ball cursor.

[Florent Daudens @fdaudens]

People are throwing "AI slop" at everything now. New study checked 25M comments on Reddit + HN: the accusations grew 60x… and accusers are constantly wrong.

The prose features that actually distinguish AI from human writing? They don't predict who gets accused.
arxiv 2606.12073

[Florent Daudens @fdaudens]

À CLI for Colab

[Omar Sanseviero @osanseviero]

Super excited about the release of the Colab CLI! 🔥

Enabling your agents to easily interact with Colab is about to unlock lots of fun experiments

https://t.co/xyQ6GFINtg

[Florent Daudens @fdaudens]

New research paper on AI coding agents: ~60% of tokens are burned in code review alone, and ~54% of all tokens are just inputs (reading/re‑reading context)

In other words: today’s agentic dev stacks are paying a huge “communication tax.”

Arxiv: 2601.14470

[Florent Daudens @fdaudens]

@grinich @mastra super interesting talk. From the lens I’m focusing on -information and AI - I think it’s crucial to understand for the future.

[Florent Daudens @fdaudens]

@Teknium @NousResearch Artemis II stick

[Florent Daudens @fdaudens]

@Teknium Thanks!

[Florent Daudens @fdaudens]

@Elon_Orbit @Teknium thanks! ended up updating config.yaml and .env

[Florent Daudens @fdaudens]

OpenClaw, Anthropic version. Basically: assign work from anywhere, come back to finished results.

This is what "AI assistant" was always supposed to mean.

[Florent Daudens @fdaudens]

Absolutely loving the new sub-agents feature with Codex 🤯

Feels like managing a tiny team of thinkers… Plato waiting, Lorentz thinking, Mendel chilling… Who decided to name them after legendary minds? 😂

[Florent Daudens @fdaudens]

@BraceSproul Love the new logo btw!

[Florent Daudens @fdaudens]

Paper: https://t.co/zPnqj51Hge

[Florent Daudens @fdaudens]

From a handful of comments, AI can now figure out who you are. Fully automated. At scale.

New study shows that LLM agents matched 67% of pseudonymous HN accounts to real LinkedIn profiles (90% precision). Best non-LLM method: near 0%.

Pseudonymity is no longer a shield.

[Florent Daudens @fdaudens]

@rseroter @addyosmani @theo I’ll definitely rewrite my AGENTS.md file!