CTO & cofondateur de https://nitter.cf/t.co/fBdEU8XbjQ & https://nitter.cf/t.co/ohuKRlYnYm
Aix en Provence
Joined August 2021
- Tweets1.6K
- Following109
- Followers114
- Likes630
Thomas Mahier retweeted
Are Agents soon going to have only 1 tool? We went from a handful of tools, to dynamic tool search, and now to collapsing everything into code mode.
Wrote down how Code Mode works, why a typed JS/Python sandbox can beat raw bash, and why MCP is going to go back strong in 2027. With an a minimal example for Gemini 4 Argon.
philschmid.de/code-mode
If you plan to explore code mode i can recommend testing with pi! Easily hackable and comes with strong foundation.
Thomas Mahier retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Thomas Mahier retweeted
When you think about the Labs' weird approach to software (sudden changes, bad communication), it is worth remembering that a core belief is that these products are a temporary thing they have to do to raise money for a couple years until "true" AGI makes all products obsolete.
Thomas Mahier retweeted
Vibecoders, 𝗜'𝗺 𝗯𝗲𝗴𝗴𝗶𝗻𝗴 𝘆𝗼𝘂, before your next website or app, paste this into Claude Code or Codex:
"I want this to be as perfect as Linear's website or Apple Notes: simple purpose, simple architecture, simple design, simple implementation. Cut taglines, helper text and secondary labels. If a screen needs them to make sense, fix the layout. Every decision should elevate the user experience (UX), developer experience (DX) and agent experience (AX) at once, without breaking anything."
Thomas Mahier retweeted
So @Ubuntu still hasn't patched this open hole on any Ubuntu server on the public internet that lets anyone take it over
We might just have to switch all our servers over to Debian in this new post-AGI time of rapid cybersecurity hacks which require just as rapid patching
Slowpokes will be left behind
chalmovsky.com/posts/wtf-ubu…
This is why I said you want your VPS ports all closed
Then only let it be accessible inside your Tailscale for SSH and via a Cloudflare Tunnel to receive web traffic
Any Linux box can be taken over with just a single TCP port exposed to the internet now
A stop gap is disabling MPTCP on your VPS now but even better is not having any public ports exposed!
Thomas Mahier retweeted
If you're running Ubuntu or Red Hat on a public server and cannot patch this yet the solution is:
sudo sysctl -w net.mptcp.enabled=0
‼️ Ubuntu and Red Hat still have no fix for a Linux kernel flaw that Anthropic says its unreleased AI model turned into a remote root exploit.
The bug, CVE-2026-98432, is a use-after-free in the IPv4 TCP handshake, rated 9.8 out of 10.
The exploit needs no credentials and targets servers with a multipath TCP listener, which Go 1.24 and later open by default.
The fix reached stable kernels on 11 October and Debian has shipped it. Ubuntu still marks every release from 16.04 to 26.10 vulnerable.
Red Hat now rates it Critical but has no patched kernel for RHEL 8, 9 or 10.
Anthropic has not released the exploit. By its count, it takes 48 minutes to leak a kernel pointer and five more to run commands as root.
Thomas Mahier retweeted
Validators! CVE-2026-98432 is patched in the linux kernel but not yet patched on ubuntu. Here's what you need to do if you're running ubuntu:
Thomas Mahier retweeted
A CVE (Common Vulnerabilities and Exposures) is a publicly disclosed security flaw.
They increased following a slight upward slope all the way to early 2024, when there was a sudden small increase... then starting in 2026, we went full exponential.
In most instances, the vulnerabilities were pre-existing, so much of the increase is just our better ability to find the faults.
This is of historical importance. We have never witnessed anything of the sort.
Thomas Mahier retweeted
When I first told @MParakhin that we were planning to move to a monorepo he said wisely:
Moving to a Monorepo is an idea that every engineer absolutely hates, and that once its there, no engineer would ever want to work at a company that doesn't use one. We will definitely do it.
Thomas Mahier retweeted
What a disaster atlasum.net/blog/cve-2026-98…
Thomas Mahier retweeted
Hard to overstate how big of a deal this is.
tl;dr — ANY unpatched Linux server with even a single TCP port exposed to the internet (say for HTTP or SSH) and MPTCP active (default for many Go web servers) is vulnerable to complete takeover.
What a disaster atlasum.net/blog/cve-2026-98…
Thomas Mahier retweeted
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5:
hackertube.org/
- Excellent recs without any logins (state is kept in local browser session)
- Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads
- Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version
- HackerTube Longs ✊
A bit more on the process:
I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing.
Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.
Thomas Mahier retweeted
This is my favorite essay I've read in the past 3 years or so. It argues that what all we and all other lifeforms really want is "Adaptive Complexity": a sort of Goldilocks-like perfect balance of complexity in our lives: not too stressed, not too bored. It's easy to pick it apart, but it's a useful framing. You can use it to explain all sorts of phenomena, such as why toxic fan bases form. Enjoyable read, highly recommend.
contrapositive.substack.com/…
Thomas Mahier retweeted
dissapointed in myself, just discovered cloudflare.config.ts
developers.cloudflare.com/cf…
Thomas Mahier retweeted
I'm convinced that at this point, you could have an Opus 5.5 agent re-implement all of CPython probably within two weeks without any human input, against just the tests, and probably with full compatibility with including ABI, no GIL, a jit and probably beat it on all measurable metrics. It's completely crazy.
Thomas Mahier retweeted
I cannot tell you how bullish I am on Stripe in the agentic era.
They are the ONLY connector inside Grok Bot that have the kind of granular permissions + human-in-the-loop prompts I *expect* for agents.
Every other connector - like ones from Google for eg Gmail or Docs - feel hopelessly amateur. Like WHY do I not have a way to have the Gmail connector ask for this kind of user permission before it lets an agent read any *sensitive* emails like password reset ones etc.
Stripe ygmi and I have no idea why others are not copying the very least what Stripe does
Thomas Mahier retweeted
"We're in humanity's golden age." I'll likely be fine whatever happens. But watching my kids grow up next to a war and with a society that doesn't know how to finance itself, let alone work with AI, I sometimes struggle to see the gold. dark.ronacher.eu/2026/10/31/…
Thomas Mahier retweeted
This is f*cking insane. Anthropic Research Lead:
"Prompting is basically over
99% of our engineers already run swarms of 300+ self-improving agents. Now everyone is shifting to graphs"
In just 20 minutes, she explains how graph engineering turns separate agents into a system that improves itself
One Agent → Loops → Plan Mode → Graphs → Self-Improving System
Prompting was the old workflow. The graph around the model is the new one
Worth more than most $1,000 agent engineering courses
Bookmark and watch it today
Then read the article below ↓
