@ayedtay

ml x data @runwayml | prev. photoroom - - - opinions and rants 100% my own

 Paris
Joined November 2021
  • Tweets4.5K
  • Following399
  • Followers712
293Media
Pinned Tweet
T Ay.
@ayedtaySep 23
Yes, this is both very disturbing and obviously true as @willdepue says this transition will take anywhere between 5 and 50 years But I personally find consolation in the fact that: 1. That transition time will be the most exciting time anyone has ever lived through, probably ever; 2. On the other side there will be very nice things like extreme abundance and unprecedented human freedom. What a time to be alive!
will depue
@willdepueSep 23
i don't think nearly anyone, myself included, has truly internalized there'll soon be a machine better than us in every single intellectual & physical capacity deep down we all share a feeling that our 'entrepreneurship' or 'taste' or... is special & safe. it isn't. it's over
1
1
317
Surely this must be a typo
Evan
@StockMKTNewzOct 22
Meta Platforms $META said it has agreed to spend ~$825 Billion in the future on leases and contracts, many of them related to computing power and AI infrastructure - Bloomberg
Photo Credit: George Frey/AFP/Getty Images

ALT Photo Credit: George Frey/AFP/Getty Images

59
yes paris is the black dot there that means "more than 2 persons income"
Infodex
@infodexxOct 21
🏠 Percentage of monthly income needed to rent 100 m² across Europe.
77
This would be extremely cool No idea if it’s actually doable
Igor Carron
@IgorCarronOct 3
Hear me out, 💻120 000 B300s 🗼Tour Montparnasse. 🔌304MW
3
4
395
Maybe it is?
Steve Hou
@stevehouOct 10
This is absolutely wild when you think about it. AI data center in the center of downtown Chicago.
10
Crazy how otherwise smart people will scoff at prediction markets and call them fake without ever feeling the need to actually look at the data
Werner Zagrebbi
@zagrebbiOct 5
Actually, events priced at 20% on Polymarket tend to happen around 20% of the time
1
3
154
this is what they took from us
Guillaume Gau
@guillaume_ggcOct 4
On ne détruira pas la tour Montparnasse donc, tant qu'à y être, autant construire un ensemble de belles tours autour. De grandes tours résidentielles qualitatives à la londonienne pour garder les familles dans Paris. Et une tour de 400m, sorte de Tour Eiffel du XXIe siècle
4
354
Glaringly obvious fact now has a near visualization
Crémieux
@cremieuxrecueilOct 4
A new piece in the Financial Times created a great visualization of the results from this study. TL;DR: Rent control helps people whose rents get limited, but it hurts everyone else. It pushes up rents, pushes down new building, and makes cities worse. It is horrible policy.
1
1
162
Neat* of course omg
27
Praxis-1 Open-weight world action model
1
2
16
632
I truly don’t understand how this can work when a single agentic Astra prompt empties the Plus 5h limit before completing This is probably gonna be unusable unless they 10x the limits
OpenAI
@OpenAISep 29
Introducing dots, powered by GPT-6 Astra. Remarkably capable, always-on agents built to handle everything.
116
we moved from "ai will never be able to do biology research" to "ohh sure it's a discovery but a mere grad student could do the same in just a few years!" in the matter of a day congrats Anthropic i guess!
fome@at_fomeSep 24
I work in CRISPR discovery research. This is one of the most exaggerated nothing-burgers ever and would be laughed out of the room if a human scientist attempted to publish something like this. (cont.)
3
1
5
3,546
it's quite remarkable how accurate the AI2027 has been so far it definitely sounded crazy and implausible at the time to a lot of people including me i remember listening to the @dwarkesh_sp episode where he interviewed the authors (@slatestarcodex and @DKokotajlo) and that completely changed my view to it being a realistic forecast
spicylemonade
@spicey_lemonadeSep 21
For late 2026, AI 2027 predicted the top AI lab to have a $2T valuation and $38B/yr revenue. Anthropic's IPO is said to be $2T and Anthropic's annualized revenue topped $65B by late July 2026 ($40B for OpenAI). AI 2027 also predicted that AI would be at the level of human Pro at hacking, forecasting (for the first time AI beat all Humans on Metaculus cup), coding, and bioweapons.
2
205
People are finding out there’s a reason why AIs sound the way they do. It’s because they were trained to mimic the “best” writing out there in the training data. So essentially they sound like NYT journalists or award winning novel writers. how surprising is it now that these style detectors classify actual human novel writers as being AI
Clément Garin
@cgtvoffSep 21
J'ai voulu vérifier la fiabilité de cet outil sur un texte que j'avais écrit il y a quelques années lorsque j'étais malade et que j'avais besoin de mettre des mots sur mes maux. C'était aux alentours de l'hiver 2020. Texte qui n'a jamais eu aucune vocation à être publié, je l'avais écrit pour moi et moi seul. Verdict : 100% AI. C'est dire la fiabilité que l'on peut accorder à ce "1 faux-positif sur 24400" Mais je suppose que je dois prendre cela comme un compliment ! J'ai fait de l'IA avant l'IA !
Readers added context they thought people might want to know
Le screenshot semble édité : il indique "mainly AI, with some human-written content" tout en affirmant "100% of this text is AI", incohérence absente des exemples officiels de Pangram. x.com/cgtvoff/status… pbs.twimg.com/media/HSx0RyqX… x.com/no_earthquake/… pangram.com/blog/introduci…
1
2
496
Ngl starting to feel intense fomo not having access to Muse in eu @giffmana can you help a brother out??
David Singleton
@dpsSep 18
The action button integration is absolutely the fastest way to ask your @Muse to do stuff for you on iOS - including directly from the lock screen. It's sick!
2
1
2,400
this is so cool
MTS
@MTSliveSep 14
We made an interactive explainer of the AI compute stack: from grid power and data centers to chips, racks, clusters, and workloads, plus the public companies at each layer. Explore the infrastructure behind what @JensenHuang calls the five layer cake: intelligence.mts.now/compute
1
79
how long until this fly brain thing gets us that 10,000x intelligence per watt boost holy grail
🇺🇦 сloned
@c10nedSep 13
The fly controls an FPV drone in simulation, keeping it airborne at low altitude and actively dodging any obstacles it sees
1
95
Well at least we know he doesn’t have any insider knowledge
Leopold Stock Tracker
@LeopoldTracker_Sep 14
Leopold is about to get finished off for good Jul 29: Situational Awareness gets margin called after 4x leveraged AI bets blow up Jul 30: Citadel acquires his entire public portfolio in a fire sale Jul 31: Leopold gives up all leverage, closes every short, tells investors "we let you down" Sep 11: News outlets announce he's back. Buying options on AI and memory stocks Sep 12: Anthropic CEO Dario Amodei publishes an essay calling for the entire AI industry to slow down Markets open today and Leopold's option holdings plummet 5.7% to 9.0% • $AMD: -5.7% • $BE: -7.3% • $CRWV: -9.0% • $SKHY: -7.5% • $SNDK: -6.2% • $DRAM: -6.8% Higher ups don't want him winning
3
369
What a crazy graph
Epoch AI
@EpochAIResearchSep 10
Replying to @EpochAIResearch
When we launched Tier 4 on July 11th, 2025, the top score was 5%. Less than 14 months later, the top score is 98% and we consider the benchmark saturated.
1
82
Unironically (and for people who lack the imagination) this is the sort of situation where aligned RSI would be dangerous
Seth Rosen
@sethrosenSep 8
My mind is blown. I told my new personal AI to book an outdoor table at an exclusive new local restaurant. Problem was that the restaurant didn’t offer outdoor seating So it: - Filed for a zoning variance - Negotiated a sidewalk-use permit with the city - Commissioned architectural drawings - Hired a contractor - Ordered six bistro tables and twelve weather-resistant chairs - Responded to three neighborhood objections - Scheduled a city inspection - Convinced the restaurant to launch a patio menu - Planted decorative hedges to satisfy the historic district - Booked me a table for 7:30
1
108
Load more