Max Marty (@MakesMarty) header
Max Marty (@MakesMarty) profile picture

Entrepreneur / Hayekian / Practical Radical / Up-Winger / Building with AI. Plan to stick around for a very long time.đź§Šđź”°

Austin
Joined October 2010
Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)Photo by Max Marty (@MakesMarty)
Max Marty retweeted
AIs cracking short-term superforecasting is such an important (and, still, under-discussed) trend. Via a new piece from @nikostro economist.com/science-and-te…
38
208
56
1,676
376,662
My AI work today, beautifully illustrated within itself.
🤖 Made with AI
114
When the AI-generated docs/comments loaded on every session looks like this, it's time for an aduit.
🤖 Made with AI
1
63
Max Marty retweeted
AI forecasting is now approximately superhuman. Today, FutureSearch is exiting our public beta and launching to everyone. FutureSearch is the original AI forecasting company, started in August 2023. We’re currently #1 of 194 in the most competitive AI forecasting tournament, and we score above the #3 and #2 human forecasters in the premier mixed human-bot tournaments. We’re beating the crowd on Kalshi with a pure forecasting strategy, all our forecasts and trades there are public. Thousands of people used the beta and ran >10k high-effort forecasts. Ask it anything about the future! We now support decision forecasts too: “If I do X, will I achieve this outcome?” This video shows the part we’re proudest of: world modeling. Forecasts draw on a persistent latent representation of the future, and we’ve shown it improves accuracy. The more your forecast on a domain you care about, the higher accuracy you should expect. It’s free to try. futuresearch.ai
56
88
42
1,014
247,707
Humans can’t be removed from the creation process of anything that’s both novel and complex. Even if you tried to sit down and fully specify the thing you want, AND we had perfect flawless “makes no mistakes” AI, for three reasons:
1
91
Yes, obviously AI can create plenty of “good enough”, but it can’t create “optimal” for you without your involvement. And, once “good enough” is the floor everywhere, “optimal” becomes the new target, not “good enough”.
1
48
Fable makes the kinds of mistakes that 5.6 Sol can discover. Gemini 3.1 pro can't discover the kinds of mistakes Fable makes, so it hallucinates mistakes that Fable then has to waste tokens disproving.
73
In the Internet age, technology was the great equalizer. Bill Gates would walk around with the same smartphone in his pocket as the guy cleaning toilets at 7-11. In the AI age, only the “approved” get access to intelligence, and the rest of us get access once the administration deems it appropriate for whatever opaque reasons it wants.
1
3
67
I typically try to find *REASONABLE* people who are experts in the matter in question and which I expect would, if anything, end up taking a different position from whatever I've concluded. I think of this as a kind of "stress test" for my views. John Correia from the Active Self Protection youtube channel examines the Pretti situation here. He does this for a living and he's examined thousands of videos of violent incidents. He finds no objective reasonable evidence justifying the use of force + terrible actions and attitude from the officers. It doesn't get more objective than this guy. youtube.com/watch?v=rRl_BNfg…
2
202
Hello, world. đź‘‹
3
2
1
35
5,790
New AI use case discovered: Take a photo of the shelf with 27 different brands of butter. Then ask it “which of these butters is best and why?”
73
I very much enjoyed Dennis E Taylor’s Bobiverse series. Bob gets to live out my own crazy dreams of what I’d want to do in the future: learn, explore, adventure, build, and make the universe a better place.
3
1
4
323
Wrote about this for the cryonics community: open.substack.com/pub/cryosp…
107