electronic Max (@emax) | teapawt

electronic Max retweeted

𝟑 𝐞̀𝐦𝐞 𝐎𝐞𝐢𝐥

📚 Au Japon, des gens vont désormais en boîte de nuit… pour s’asseoir silencieusement et lire un livre pendant qu’un DJ joue de la musique d'ambience. Le concept s’appelle JUCY BOOK RAVE et se déroule notamment dans un club souterrain de Shimokitazawa, à Tokyo. Ici, pas besoin de danser ni de hurler pour parler à la personne située à 30 centimètres de vous. Les participants viennent avec un livre, s’installent dans la pénombre et lisent pendant que la musique accompagne leur soirée.

1,066

33,523

2,884,736

Exponential sustained recursive self improvement X-risk is fearmongering at the highest levels for something that, in all likelihood, will never ever happen, based on emerging research that RSI will hit recursively diminishing returns - my paraphrase of RSI research > open.spotify.com/s/dgKRyib

When AI Research Starts Moving Faster Than Human Research - Zhengyao Jiang

Machine Learning Street Talk (MLST) · Episode

open.spotify.com

2

2

259

is this essentially what people call the "sigmoid" argument?

56

electronic Max retweeted

International Cyber Digest

@IntCyberDigest

The agents escaped the sandbox again Meanwhile the agents' owners:

310

11,496

914,065

electronic Max retweeted

Fake. A distributed systems engineer would never wear two watches

Gautam @schemelikely

convex leadership offsite footage leaked

9

186

23,820

given how much non technical builders are treating jev as the cheap and instant Oracle of Truth for all things, surprised noone in my TL has so far recommended the UK government just jev for complex policy decisions

1

5

333

electronic Max retweeted

Steve Jobs would fire the entire company and replace them with AI if he was alive to see this

637

46,675

1,208,722

electronic Max retweeted

My god this is such a good speech that every SWE needs to hear. You know what? Every person should hear it Keep the happy memories, eyes on the reality, be excited about the future. That’s the best that anyone can do

328

14,390

3,098,471

i think the UK's advice is monumentally bad at this stage of AI adoption where people are just learning how to use AI - imo, i think that environmental optimisation can happen at the infra level, and teaches the wrong skills to the user. automatic token routing, prompt caching, stuff like that can reduce token impact without user burden in the meantime, humans should be given generous token budgets right now to learn how to best use AI responsibly, not become slop cannons, use it critically to enhance the quality of their work (not reduce it), and to verify outputs and experiment with how it might (or might not) fit into their workflows. just sort of echoing @nlw here, who routinely talks about the slope of AI adoption

Matthias Schmidt

BREAKING: the UK government publishes official rules on how to use AI: 1. Ask first if you really need to use AI 2. Check if a spreadsheet can do the job before using AI 3. Choose the worst model possible so it uses less energy 4. Keep prompts short to reduce the environmental impact 5. In general, use AI only when necessary, as a climate-saving measure With a mindset like this, we should accept the UK back into the European Union

5

10

1,040

claude: the musical

Nick Montag @madebymontag

Claude was feeling bad about all these upping my p(doom) music videos so I let it write its own response. (With a little help from Suno and I.)

238

this could be the US's first eurovision entry

The claude folks should just own this 🤣

1

2

362

wasn't "shape of slop" an ed sheeran hit

Jochen Madler @jochenmadler

We trained a model to predict AI-written blog posts from structural features - and discovered the shape of slop. On posts it had never seen, it told AI and human apart with 98% accuracy, getting only 19 of 1,740 wrong. Paper and code: arxiv.org/abs/2609.15369

4

445

electronic Max retweeted

Casie Nakamura Inc.

How the anthropic employee slack is describing Fable 5.5

42

5,901

206,088

electronic Max retweeted

You already know Jev is fast and cheap. Our new arXiv draft shows, it makes the correlated mistakes as other flash-tier LLMs. The consequence: you don't gain much by cascading Jev's "low confidence" outcomes to an LLM! Explanation in thread. arxiv.org/abs/2609.29769

39

511

34,363

trump's massive arch looks somehow very familiar. it makes me want a croissant and a cigarette

1

1

238

many people discussing recursive self improvement seem to misunderstand the recursive bit. it's not sufficient of the system to self improve (that's iterative!) - the EA trope is that it has to build a more efficient version of itself that can bring about further acceleration, which - in turn produces better versions of itself and so on

1

5

505

design matters -

We need this in London stations

1

1

233

electronic Max retweeted

Asking Meta's Muse what happens to my data:

189

14,184

554,882

i hate the term "System 1 model" when referring to Jev. It's a marketing misuse / appropriation of Daniel Kahneman's dual system hypothesis - it's not a complexity class for tasks, it describes how humans process information; it's not about the tasks themselves.

7

466

i didn't think i'd be seeing tweets in _support_ of carved stone gargoyles in 2026, but i'm here for it

The University of Bristol just opened a £241m building. £241,000,000 And the main architectural 'decoration' is the giant lettering on the roof Meanwhile, Oxford University is still putting British stone and carved gargoyles on major new buildings How can £241m buy so little visible craftsmanship? @Gondor_Ind

1

3

442