Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile
Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค

@flolightc

๐Ÿ’ปPrototype Architect @awscloud
๐ŸŒฅCloud and AIML / GenAI enthusiast
๐ŸŽšStage lighting passionate

๐Ÿ“ŒOpinions here my own โ˜„
he / him
Follow me to grow together!

ID: 1369152438

linkhttps://bio.link/florianclanet calendar_today21-04-2013 09:50:10

6,6K Tweet

2,2K Followers

3,3K Following

โ„ฮตsam (@hesamation) 's Twitter Profile Photo

just finished the Reasoning course on one sit (takes ~1 hour), it's a really cool and simple introduction to: > reasoning in LLMs > reinforcement learning in LLMs > the deepseek r1 paper and GRPO > practical example of grpo fine-tuning more units will be added to the course too.

just finished the Reasoning course on one sit (takes ~1 hour), it's a really cool and simple introduction to:
> reasoning in LLMs
> reinforcement learning in LLMs
> the deepseek r1 paper and GRPO
> practical example of grpo fine-tuning
more units will be added to the course too.
Baptiste Robert (@fs0c131y) 's Twitter Profile Photo

Hi Elon Musk, I've identified the people responsible for the DDoS attack on X yesterday. I'm currently in Washington and will be at the Eisenhower Building tomorrow (for another matter). Would you be interested in meeting? In the meantime, let me explain It's OSINT time!

Hi <a href="/elonmusk/">Elon Musk</a>,

I've identified the people responsible for the DDoS attack on X yesterday. I'm currently in Washington and will be at the Eisenhower Building tomorrow (for another matter). Would you be interested in meeting?

In the meantime, let me explain

It's OSINT time!
Cline (@cline) 's Twitter Profile Photo

Memory Bank isn't about minimizing tokens -- it's about maximizing project efficiency. Yes, reading memory files costs tokens. Cline users confirm this. But they also report: >More context = fewer hallucinations >Fewer hallucinations = less rework >Less rework = fewer total

Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile Photo

Read โ€œChain-of-Draft (CoD) Is The New King Of Prompting Techniquesโ€œ by Dr. Ashish Bamania on Medium: levelup.gitconnected.com/chain-of-draftโ€ฆ

โ„ฮตsam (@hesamation) 's Twitter Profile Photo

this 211-page book "Pen and Paper exercises in machine learning" is available for free on arxiv, a based way to learn ML math and theory: > linear algebra problems and solutions > vector and matrix gradients > directed & undirected graph models > maximum likelihood estimation

this 211-page book "Pen and Paper exercises in machine learning" is available for free on arxiv, 
a based way to learn ML math and theory:
&gt; linear algebra problems and solutions
&gt; vector and matrix gradients
&gt; directed &amp; undirected graph models
&gt; maximum likelihood estimation
โ„ฮตsam (@hesamation) 's Twitter Profile Photo

learn these 10 Python tools if you want to work on AI engineering projects: 1. uvย โ€“ Lightning-fast package & dependency manager. 2. ruffย โ€“ An extremely fast Python linter and code formatter. 3. AdalFlowย โ€“ a PyTorch-like library to build and auto-optimize any LLM workflows. 4.

Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile Photo

An interesting proposal to standardise on using an /llms.txt file to provide information to help LLMs use a website at inference time. github.com/answerdotai/llโ€ฆ llmstxt.org

Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile Photo

Je participe au concours avec #lfpoulain et #AsusFrance pour gagner un kit de deux Routeurs Extensibles WiFi 7 : ASUS ZenWiFi BT8 ! ร‡a se passe par ici : fr.asus.click/concours-wifi7โ€ฆ

Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile Photo

I'm exploring agentic GenAI for a few months, and I'm still finding it super interesting! It's cool to see how these systems can operate autonomously. Anyone else into agentic solutions? Would love to hear your thoughts and experiences, especially in production.

Alex Vacca (@itsalexvacca) 's Twitter Profile Photo

BREAKING: MIT just completed the first brain scan study of ChatGPT users & the results are terrifying. Turns out, AI isn't making us more productive. It's making us cognitively bankrupt. Here's what 4 months of data revealed: (hint: we've been measuring productivity all wrong)

BREAKING: MIT just completed the first brain scan study of ChatGPT users &amp; the results are terrifying.

Turns out, AI isn't making us more productive. It's making us cognitively bankrupt.

Here's what 4 months of data revealed:

(hint: we've been measuring productivity all wrong)
Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile Photo

Just stumbled on "use-mcp" repository. Super handy for linking React apps with MCP servers. Makes MCP integrations a breeze! github.com/modelcontextprโ€ฆ

Florian Clanet ๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป ๐ŸŒค (@flolightc) 's Twitter Profile Photo

๐Ÿ” Just read an interesting blog post on OAuth2 and the Model Context Protocol. It's all about making agent communication secure. It might help with the authentication question your company is wondering about with MCP go.aws/4epNGUx