
💻Prototype Architect @awscloud 🌥Cloud, GenAI/AIML enthusiast 🎚Stage lighting passionate 📌Opinions here my own ☄ he / him Follow me to grow together!
Toulouse, France
Joined April 2013
- Tweets6.6K
- Following4.1K
- Followers2.5K
- Likes25.8K
Pinned Tweet
Hey, nice to meet you, I’m Florian !
Coding enthusiast and building cool things in the Cloud⛅️
Curiosity is my main sin and I’m trying to share it with you !
I also talk about cybersecurity, lighting design and try to find out the purpose of life 🧙♂️
📖
medium.com/@florianclanet
Florian Clanet 👨🏻💻 🌤 retweeted
GIVEAWAY: We're giving away a Canaan Avalon Nano 3S.
Up to 6 TH/s of solo mining right on your desk, and Canaan just open-sourced its firmware.
To enter:
❤️ Like this post
🔁 Repost it
➕ Follow us
We ship anywhere in the world, and we cover the taxes and duties. Good luck, stackers.
Check this one !
CodeKB がOSS として公開されました。これはAI-DLC の取り組みなどからわかった既存のシステム(brownfield) の開発で必要となる既存コードのAIによる理解を助けるためのツールです。ぜひ触ってみていただき、フィードバックをお寄せください!
github.com/awslabs/codekb
1
87
Florian Clanet 👨🏻💻 🌤 retweeted
‼️ BREAKING: Researchers broke past PHP's fix for the php2shell flaw and can again run their own code on web servers, no login needed.
Security firm depthfirst says the latest PHP releases are still vulnerable. The bug, CVE-2026-126714, is rated 9.2 out of 10.
depthfirst.com/research/phpa…
Readers added context they thought people might want to know
PHP 8.4.28 and 8.5.13 fixes CVE-2026-126714.
php.net/ChangeLog-8.ph…
php.net/ChangeLog-8.ph…
Clare is an AWESOME speaker, she was an instant accept at AIEWF and her talk instantly crushed it, especially hard to stand out on this track because you're in competition with some of the biggest household names in the world.
would tune in for this AWS reInvent keynote 👀
I am so excited to lead the Developer keynote at AWS reInvent this year!
Working on the talk track atm, and we've got such a great lineup of speakers, deep dives, and practical advice
Sign up for the livestream here:
aws.amazon.com/events/reinve…
Monday, Nov 30, 7:30 pm - 8:45 pm
Florian Clanet 👨🏻💻 🌤 retweeted
AWS quietly removed a ton of undifferentiated heavy lifting in running secure agent sandboxes yesterday and I don't think people outside my normal circle are aware of how useful this is going to be: docs.aws.amazon.com/AWSEC2/l…
Indeed! YOU MUST TRY THIS!
i'm sorry i dont care which model made this but it blows my mind that this is made by an llm AND runs in the browser
claude.ai/artifact/BGY1SztrA…
80
Florian Clanet 👨🏻💻 🌤 retweeted
We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄
To enter:
🟢 Share this post
🟢 Comment #RTXON
Florian Clanet 👨🏻💻 🌤 retweeted
We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄
To enter:
🟢 Share this post
🟢 Comment #RTXON
Terms and Conditions: nvidia.com/en-us/geforce/con…
Some news: I’m joining OpenAI to lead product management for Core Products!
I have a lot of respect for OpenAI for showing the world that scaling works, for bringing AI to over a billion people, and for believing great products are core to the mission.
I’m excited to work with @thsottiaux, an incredible PM team, and everyone across OpenAI. Together, we’ll build agents that have your back and your team’s, and keep you in control, so you get more time for the people and work you love.
Lots to learn, and lots to build. Please keep your feedback coming!
Interesting discovery with @lunaroute_ai 🤯
I'm still not completely sure how it works behind but I'll try it for sure...
106
Florian Clanet 👨🏻💻 🌤 retweeted
Do individual limits still make sense? As capabilities increase you want to work with and support an agent that represents your project, that runs loops and evolves your codebase. The humans merely curate the automations.
NEW: Apple is cutting employee AI spending budgets even as it pushes workers to use the technology more. Engineers have seen daily limits cut from $300 to $200, with further reductions to $100 planned. Other teams could see limits fall to $25. bloomberg.com/news/articles/…
I'm speaking at #reInvent in a Builders' Session on live fact-checking.
My colleagues and I built a prototype that checks claims during a live TV debate and gives a verdict (true, misleading or false) while the debate goes on.
In IND421, you build your own chain of fact-checking agents on Amazon Bedrock AgentCore, in 60 minutes.
💻 Bring your laptop
📅 Tue Dec 1, 4:30 PM, MGM
📅 Thu Dec 3, 10 AM, Caesars Forum
Link bellow 👇
1
71
Florian Clanet 👨🏻💻 🌤 retweeted
‼️ Anthropic and Google researchers found a flaw that lets one crafted packet hijack a server during network boot, before Secure Boot checks the OS.
Almost all servers on earth are affected.
CVE-2026-97631 sits in TianoCore EDK II, code in the UEFI firmware of most servers.
Readers added context they thought people might want to know
The vulnerability is limited to servers actively PXE network-booting with vulnerable EDK II firmware; most servers boot from local disks and are unaffected in normal use, with Intel noting limited impact on supported products.
github.com/tianocore/edk2…
arstechnica.com/security/2026/…
Florian Clanet 👨🏻💻 🌤 retweeted
We've revamped Hermes Agent subscriptions.
Every new plan now includes a cloud agent out of the box, and usage has daily and weekly limits instead of one monthly block of credits.
Subscribers can switch to the new plans any time at portal.nousresearch.com/bill…
an update on the openai/codex ban:
i was curious about my tv, found a vuln (not with codex), popped a shell, and started poking around. the intrigue of my hobby project while i had the flu wasn't something i could've predicted.
to be clear: the ban almost certainly had nothing to do with using openai -specifically- for my samsung reverse engineering work. i have no reason to believe it did.
the real issue seems deeper. i talked to friends using astra/6.1 sol for both cyber work and normal software tasks. the ones who never hit daybreak (and weren't in the blue program) said they spend a while "warming it up", getting the agent into a flowstate with substitute words and roundabout explanations that jailbreak the context. once that's done, the agent starts talking the same way and skipping tokens.
if you do serious reverse engineering without those jailbreaks, daybreak's content-hidden tripwire looks at the prompt or output in isolation and sees something indistinguishable from malware development.
i think the better fix is making the system track intent across the whole context (and through compressions) instead of punishing blunt technical language.
i believe i got banned because i just used normal terms. talking about memory, trampoline hooks, encrypted/read-protected pages, even entirely locally, looks sinister out of context. once daybreak trips, it poisons the context cache. new conversations start failing too. eventually even normal tooling work and /goals would randomly trip it. arguing your case makes it worse.
the incentives here seem strange to me. people who obfuscate from the start never seem to have problems. people who just use the actual terminology get more and more refusals.
account's still closed. pretty sure it's just from tripping daybreak too many times (even on non-cyber work).
for now i'm moving to open weight models and locking down my own compute. if these tools are part of how you work now, it's worth thinking about how to keep reliable access.
please don't harass tibo, roon or other oai people. from their public comments it sounds like they already know and agree the current setup is too strict.
sorry for the interlude, brb, back to getting the report and a repo out to fix TVs.
Florian Clanet 👨🏻💻 🌤 retweeted
Microsoft propose un départ volontaire à environ 60 000 salariés
Le programme concerne certaines équipes aux États-Unis, au Canada, en Inde et au Royaume-Uni, jusqu’au niveau hiérarchique 70. Les intéressés auront jusqu’au début décembre pour choisir une sortie en fin d’année.
businessinsider.com/microsof…
Florian Clanet 👨🏻💻 🌤 retweeted
Information exclusive Blast : le Premier ministre a préparé la coupure des réseaux sociaux pour le 24 octobre
Deux jours avant la manifestation du 24 octobre, le Premier ministre Sébastien Lecornu a demandé à Orange, SFR, Free et Bouygues Telecom de se tenir prêts à couper l’accès à TikTok, Instagram ,Snapchat et X. Blast révèle la lettre et le compte rendu de la réunion qui a suivi, où les opérateurs ont exprimé leurs doutes. Le gouvernement avait déjà utilisée cette mesure liberticide en Nouvelle-Calédonie en 2024, avant d’être désavoué par le Conseil d’État.
Par @yanmdi.
blast-info.fr/articles/2026/…



























