Sabitlenmiş Tweet
Anoni Mouse
268 posts
Anoni Mouse
@__AnoniMouse__
Aging corporate engineer dreaming of building something I believe in one day. Linux🐧 NixOS❄️ Currently starring in the 3rd film of the My Life trilogy.
The Internet, Planet Earth Katılım Kasım 2025
209 Takip Edilen10 Takipçiler
Anoni Mouse retweetledi
NEW There's no overstating how significant the Nathan Gill case is. On Friday, the ex-Ukip/Reform politician and close associate of Nigel Farage was jailed for over ten years for taking multiple bribes via a handler from a friend of Putin. It's a complex, detail-heavy story..🧵⬇️
English
Anoni Mouse retweetledi
So listen.
@thenerve_news has the receipts.
A comprehensive timeline of Nathan Gill & Nigel Farage’s pro-Kremlin influencing activities & 2 other MEPs, David Cobourn & Jonathan Arnott.
Please read & share.
THE NERVE@thenerve_news
NEW There's no overstating how significant the Nathan Gill case is. On Friday, the ex-Ukip/Reform politician and close associate of Nigel Farage was jailed for over ten years for taking multiple bribes via a handler from a friend of Putin. It's a complex, detail-heavy story..🧵⬇️
English
Anoni Mouse retweetledi
My colleagues & I have taken a huge gamble to set up @thenerve_news. We’re trying to build a new independent publication from the ground up. Social media is our only distribution for now.
Sharing this article in your networks would make a huge difference. Thank you! 🙏🙏🙏
Carole Cadwalladr@carolecadwalla
NEW: The British politician, his Russian intelligence handler & a Kremlin plot against the US & Ukraine. My new piece about Nathan Gill and Nigel Farage for @thenerve_news in which we ask: Why, even now, is no-one asking questions?
English
I'm a UK remote worker - looking to connect with more engineers. Please add me 🙏
In my personal life, I am currently working on agent harnesses and coding agent benchmarks. I need data to decide between different models and approaches - vibes is not enough! Might try and open source some of it if I feel brave.
Also about to release a tarot card app - if I could only find my way through the Google Play admin maze lol
English
Calling all London/Ukmaxxing 🇬🇧builders/engineers/developers, let's grow this list.
Comment if you want to be added
x.com/i/lists/205269…
English
@mitsuhiko @antirez I'm really excited about these projects. I will be watching closely and I will try to contribute in some small way once I get a new machine.
Local AI has to be our future. We have to crack this problem.
English
I think @antirez ds4.c is important! I wrote down my thoughts on why I built pi-ds4 and why we need to focus our local model efforts stronger than we do currently. lucumr.pocoo.org/2026/5/8/local…
English
@LLMJunky @dillon_mulroy This isn't part of the open standard though, right?
English
i think skills are a mistake and the wrong abstraction. i almost never want my agent auto invoking them and i have built custom tooling to "toggle" them on/off to prevent them from always being present in my context window.
English
why? why shouldn't i be an absolute moron and buy this?
English
I haven't really experienced this directly where I work. We have some pretty tough bugs to solve at spread, spread across a huge microservices architecture with lots of technical debt and poorly understood code. I get frustrated sometimes I'm often the only person who can follow a problem through to resolution - but I don't see a problem when a bug ticket is opened by an engineer. I can't say the same for our QA team though unfortunately :/
English
the absolute worst part about the state of our industry is that excellent engineers are now entirely incapable of formulizing useful bug reports.
we are at the level of "my computer doesn't turn on".
wtf are we even doing.
English
@strikerglows @GaryMarcus There's huge demand for heavily subsidized inference. Anthropic is making a loss. What we don't know is what demand looks like when venture capital dries up and the full cost charged.
English
@GaryMarcus The fact that Anthropic was immediately able to take over the entire capacity is literally a clear sign of enormous demand. The demand just wasn't for grok
English
just imagine what might happen if everybody was stockpiling GPUs at high prices for a demand that didn’t quite materialize….
Kakashii@kakashiii111
So, look surprised: xAI, which is now part of SpaceX, stockpiled too many Nvidia GPUs and had no use for them, because demand mostly goes to Claude, ChatGPT, and Gemini. Grok doesn't need much compute. Stockpiling out of FOMO? Oversupply? No shortage in actual GPUs? Another validation of my work.
English
@GaryMarcus If you're right, does this mean I can afford to buy GPUs for local compute after the correction? That would be a significant silver lining.
English
@0xSero Perhaps not technically a bubble, but I think there's a crash coming.
English
It’s not a bubble.
But keep thinking that, more for me.
roon@tszzl
not enough people are emotionally prepared for if it’s not a bubble
English
@ShimazuSystems I have some personal benchmarks I am running. Have you tried the new deepseek v4 pro? In my tests, I'm getting very near Opus/5.4 performance, for a tiny fraction of the price.
If you haven't checked it out yet, do take a look. I'd be interested to get your take.
English
Local/OpenSource models can do things
But like
It is not the same, it is like alien levels of not the same in certain tasks, so it's like
You always have to get the cutting edge, which always goes up as these companies cannot refuse to charge more as they are bound to by their investors
Whereas if say, Kimi, was the absolute SOTA and we stuck there? I think harnesses would eventually work around it and for most tasks eventually it would Become Good Enough
English
So, to be real here, I think we had a point where AI was good enough to build defensive measures against a future iteration of itself, that had a balance of effectiveness (code output quality) and lack of capability to plot & adapt (RLHF, RSI, etc).
I think we surpassed that point a while ago, and in all honesty it should have been a public service building this stuff without requiring to endlessly scale to meet what VCs and investors want.
So we get unprepared hard takeoff (we are here), with absolutely no restraints & substack articles at best for economic theory... oh my
English
@ShimazuSystems One of my kids just showed me a new operating system claude code has written for her. Written in C. A new kernel. FAT16. Boots as a VM to a shell.
It's vibe coded and I've no doubt it's an unmaintainable mess, but even so, she's 13. Crazy times.
English
There's usually a smartass every once in a while who proclaims 'oh why use AI then? Why just not use it?'
I cannot, as an engineer, as a person who thrives on creation, turn down this tool
It is like crack
Things I had no need to learn because why would I? (time is finite)
English
@badlogicgames You built an amazing piece of software. It's quite an achievement 🫶
English
Anoni Mouse retweetledi
@spactastical @TheAhmadOsman My assumption is he means being unable to afford to buy access from AI providers, like Anthropic, because it's too expensive. Those without compute are at a disadvantage and get left behind?
English
@TheAhmadOsman What do you think will be the main advantage of having your own compute (besides local AI use) from a point of generating revenue? Leasing local compute? Leasing local AI services/tokens?
English
I should add that I wouldn’t wanna be compute poor comes 2029
Bookmark this
Ahmad@TheAhmadOsman
What most don’t get is that Compute is the only real moat
English
@TheAhmadOsman Curious what you think is special about 2029?
English
@Ysqander @badlogicgames What does that actually mean?
Dario is "ahead" of anyone "contradicting" him 🤔
English
@badlogicgames So far Dario is ahead of anyone contradicting him including Elon. But sure Mario, gamble away 😂
English
dario is going to have a very bad time when investors find out that the small $100k token packages with discounts he's selling don't produce any measurable ROI for those non-software corps.
Wall Street Rollup@WallStRollup
Anthropic’s Dario Amodei warns that software incumbents who don’t pivot, are going to be blindsided, and “have a really bad time”.
English