working towards better AI futures. AI safety at @snyksec. Organizing @lisbonai_. Data and capital at @capitalfactory. Prev: @gumroad @techcrunch @oreillymedia
- Tweets 16,093
- Following 936
- Followers 134,012
- Likes 11,842
Fred Oliveira 🧠 retweeted
We’ve run the most comprehensive series of studies on expert AI forecasts over the last four years. Today, we’re sharing an interim update on our findings about the accuracy of these forecasts.
Our major findings are:
1. Experts, including top economists, computer scientists, and biologists, have dramatically underestimated AI capabilities progress each year. Superforecasters have underestimated progress to an even greater extent.
2. Experts have a more mixed forecasting track record on AI diffusion-related measures, with some major underestimates (forecasting AI revenue) and other forecasts on track to be accurate (the share of electricity used for AI).
3. Some notable cases of overestimating AI progress: how much mid-2025 AI models could help amateurs do biorisk-relevant laboratory tasks; the speed of rollout of self-driving cars.
4. It is too soon to say how forecasters have performed at predicting macro-scale impact on outcomes such as GDP growth, major AI harms, and averted deaths from disease.
More 🧵
4
87
329
44,247
we did it. @lisbonai_ 2026 is over. Too many thoughts, so I will say this: a part of me is as tired as I can be, but a larger part of me is as energized as possible.
I am heading home with my heart filled by people and my mind filled with ideas.
thank you. Build.
ALT a drone shot of the Lisbon AI venue, fundação champalimaud
12
5
82
5,866
@GuliMoreno btw, you were instrumental here too. Thank you for the connections to cracked spanish talent ❤️ Whenever you come to Portugal, I'm buying the drinks. Gracias!
1
2
304
credit where credit is due, @diogomonica is a rockstar. Great talk about the OpenAI/HF incident, by framing it around the Pirates of the Caribbean
(Pirate) hats off
2
2
33
3,009
Fred Oliveira 🧠 retweeted
Day 2 is live.
We'll have AI on the canvas, AI on the screen, AI on robot brains, AI for drug discovery, and that's just the very first block.
LFG 🤖
We're live.
Day two starts now.
Applied AI up first - turning models into products people actually use.
Watch here: nitter.net/i/broadcasts/1oKMvNkjA…
1
6
510
some people walk to the venue, some people uber, some people bike. Matt, on the other hand 👇👏
3
4
1,805
you know it is a good tech conference when you compare @obsdmd graphs with the speakers. Here's mine, if you are curious.
ALT my obsidian graph setup
2
14
1,013
@kepano you'll be happy to hear that Obsidian has been on stage many times already, and we're still only half way through day 1. A testament to great work.
1
288
if you want to follow LisbonAI and can't be in the room, we created a live page where you can look at the schedule and the stream at the same time: lisbonai.org/live/
3
16
932
insanely good start to the @lisbonai_ program.
FYI, we’re pausing the stream for the next talk because @chema_garabito is going to show some exclusive stuff, but we’ll be back afterwards. 💪
We're live.
Day one kicks off with Models block: @Prince_Canuma, @joanrod_ai and @chema_garabito, live on stage.
Watch here: nitter.net/i/broadcasts/1qJDzWdoQ…
1
2
11
2,254
if this weather and location is any indication, this year’s @lisbonai_ is going to be amazing.
2
2
31
3,350
Fred Oliveira 🧠 retweeted
Many are demanding more concrete scenarios, rigorous models etc. of how AI takes over.
When these don't meet their standards, they claim the risk is low.
That's a bad inference, but more importantly, why place the burden of proof there?
Building swarms of AI agents that are smarter, harder working, more creative, and more coordinated than humans is risky on its face.
And no-one has anything remotely like a rigorous model demonstrating that these systems will be safe either.
Optimists can't even explain how obvious problems like reward hacking and instrumental convergence will be overcome.
All they have is a broad belief that technology is good, and a hope that we can muddle through iteratively.
31
10
90
5,547
one simple trick to never have to work again
get it?
1/ Better robot models come from better data.
That takes two things: hardware that captures it fully, and tooling that turns it into training signal.
We built the first. Today, @GroundedSI launches the second: Grounded API. Access both hardware + API ↓
2
6
3,664
Fred Oliveira 🧠 retweeted
I'm giving away 2 tickets to @lisbonai_ to students. The more knee-deep in AI fundamentals, the better.
Who should get them? DM me.
@alcides @ABarroso @locks @ferreira_hugo @belchior Academia was mostly MIA last year. Let's fix that.
8
10
1,304
spot on
i've tried and failed to make this point as clearly as i want to, but:
AI is weird, in that ppl who are optimistic about the technology's growth are often doomy about its implications; while ppl who are extremely skeptical about AI's capabilities are by default optimistic about it being a "normal" technology that won't change much in the world.
so you have this weird situation where a lot of AI skeptics who haven't been following AI news for the last few years (bc they wrote the whole thing off as NFTs 2.0) have no idea that there exists an enormous contingent of AI ppl who have for 10+ years said that very fast AI progress could be really bad for the world, and therefore they suspect that a conspiracy was concocted in the last two weeks to invent the case for AI safety
3
3,860
