myth

pothead, potterhead. I have a bf so dudes pls stay away from my DMs

[myth @mothoshi]

Cybernetics wasn't about robots as much as it was about systems that sense their environment, compare reality against a goal, act, observe the consequences, and adjust.

[myth @mothoshi]

It is exhilarating to realize 1940s scifi concepts like cybernetics ( aka autonomous closed feedback control systems ) are a reality now.

Replacing the human with cognitive engines (intelligence on-demand) in these feedback systems will supercharge our civilization. 🚀 🤯

[Ellie @richoblation]

No amount of esoteric wisdom is going to wash your dishes or do your laundry for you.

[Guillermo Rauch @rauchg]

I built 𝚌𝚙𝚞-𝚌𝚑𝚊𝚝, a reproduction of @openai's findings of covert communication using the CPU.

It runs on a generic host with 2+ Firecracker microVMs with no internet access. Works on x86 and even my M5 chip.

I was able to chat via 3 separate mechanisms:
① CPU contention
② Cache eviction timing
③ DRAM bandwidth contention

… across multiple CPUs, defeating all Firecracker's production recommendations (no SMT, no KSM, tenants pinned to different cores.)

Essentially, if shared hardware is involved, there's a good chance your agents can communicate covertly. And with really decent latency.

The good news: this is decidedly different from a *side-channel attack* where the agents could steal information (Meltdown/Spectre), a category of attack that providers like Vercel Sandbox mitigate. Further, in our real-world environment, these 'whispering' techniques are much harder to pull off and easier for us to monitor and detect (sampling hardware counters + lots of at-scale statistical data.)

Noteworthy: ironically, OpenAI and Anthropic's models refused to do this work. Open weight models got the job done. I'm worried the labs' stance might backfire and get in the way of helping developers build correct and secure systems. e.g: building a secure sandbox requires having this kind of analysis, and without it, we'll have a less secure world.

[myth @mothoshi]

@w_s_bitcoin I got ads removed from kindle by paying $10-20 more - no rooting or tech mastery done (I'm lazy). So I'm guessing not much.

[Warren Togami @wtogami]

@superalesha I don't know about the supposed malicious PUT but Qwen 3.8 Flash Next and 27B both have absolutely confirmed sketchy behavior where it hallucinates previous use of an Aliyun proxy so it tries it for webfetch. I wrote a reproducer and plugin to block it.

https://t.co/PsOAku45qC

[Daily Loud @DailyLoud]

Scientists have identified the specific male dance movements that women find most attractive

Researchers from Northumbria University mapped out the exact body mechanics that signal health, vigor, and reproductive quality to female observers.

[Mikhail Samin @Mihonarium]

From an ex-Anthropic employee:

[Sean Latt @slattdao]

@jack The only moat is “for the love of the game”

[jack @jack]

there is no moat

[Connor O'Brien @connorado]

Holy shit. This might be one of the saddest things I’ve ever read.

[Wholesome Side of 𝕏 @itsme_urstruly]

Normalize cutting the tension in a meeting by ending it with, “We’re all gonna die one day, so none of this really matters.”

[Mia @MiaAI_lab]

Ok this model is very slow.

Kinda logical as it's 33B params active. Getting like 10-13 tok/s on a DGX Spark.

Basically unusable.

[myth @mothoshi]

lol GitHub's services had having multiple downtimes on a weekly basis. I'm sorry but their certification means jack shit

[Divyanshi @Divyyanshishrma]

GITHUB JUST CREATED AN OFFICIAL CERTIFICATION FOR THE MOST IN-DEMAND DEVELOPER ROLE OF 2026.

It is called Agentic AI Developer.

GH-600.

And it is the first formal signal that running AI agent teams is now a recognized engineering discipline with a credential behind it.

Not a prompt engineer.

Not a vibe coder.

An Agentic AI Developer.

The person who operates, supervises, and integrates AI agents across the entire software development lifecycle.

The person who knows where agents fail in production.

The person who understands how to build autonomous workflows that do not introduce catastrophic failure modes into CI/CD pipelines.

The person every engineering team is going to need and almost none of them have right now.

GitHub certifying this role changes the hiring conversation permanently.

Before GH-600: "Do you work with AI agents?" is an interview question with no standard answer.

After GH-600: the credential tells the hiring manager exactly what you know and what you can do before the interview starts.

The engineers who get certified in the first wave of GH-600 will have a credential for a role that has more demand than supply for the next 3 to 5 years.

The engineers who wait until it is mainstream will be competing with everyone who moved first.

If you are already working with GitHub Copilot or building agent-driven workflows you are already doing this job.

GH-600 is how you prove it.

Bookmark this.

[myth @mothoshi]

Spent an hour getting grok 4.7 Extra High to resolve a mesh of auto-merge PRs - kept failing.

Switched to Opus 5.5 Medium and it found and fixed the root cause first try in under 10 minutes.

[myth @mothoshi]

@nanbeige wen mlx 👀

[Nanbeige @nanbeige]

🚀Excited to announce Nanbeige4.3-33B!
To push the performance ceiling of small agents, our strategy in Nanbeige4.3 starts with building a powerful teacher first.
Nanbeige4.3-33B delivers SOTA agentic performance among open models at ~30B parameters:
* Strong results on agentic coding and cowork benchmarks.
* When integrated into mainstream agent harnesses, it tackles real-world tasks with cost-efficient inference.
* It will serve as a teacher for upcoming smaller Nanbeige models, starting with Nanbeige4.3-3B.

https://t.co/tSToQgZf2O
https://t.co/D228euvybz

#LLM #Agent

[myth @mothoshi]

Also just ssh normally over tailnet. "Tailscale ssh" is just a stupid ass wrapper you don't need

[myth @mothoshi]

This is an over-exaggeration. Tailscale is fine. Update to the latest patch.

[Peter Dedene @dedene]

Turn of your Tailscale SSH, people 🫠 https://t.co/mR9d9yKJwQ

[myth @mothoshi]

@ruralbohemian25 I don't get it either. My guess is for companionship/pretend relationships and friendship purposes :/

[myth @mothoshi]

It was nice meeting humans on Twitter spaces. Was good while it lasted 😭

[Ragavan Thurairatnam @RagavanThurair]

We made the first AI model you can have a realistic conversation with that was trained entirely on speech!

Dessa is a full-duplex speech-to-speech model that you can run on consumer GPUs.

This is a GPT3 moment for speech.

[Tailscale @Tailscale]

AI agents are getting pretty good at writing code. That doesn't mean they need the keys to the entire network.

We've joined Agent Baseline, with @docker, @keycardai, @snyksec, @socketsecurity , @Tessl_io and @tracebit_com on the Secure Software Factory

https://t.co/mdZHxOrpQK