Convergent sequence and entertainment @inductiv_
Exteriores Spatium
Joined January 2017
- Tweets209
- Following133
- Followers123
- Likes686
Itamar retweeted
๐ฉ When Polymarket or Kalshi says there's a 70% chance something will happen, should we believe it?
Excited to share our new paper, just published in PNAS, in which we examined more than ONE MILLION forecasts from Polymarket, Kalshi, Manifold, and Metaculus, spanning 12 domains and forecasting horizons ranging from hours to years.
The main finding is striking: Prediction markets and collective forecasting platforms are remarkably well calibrated. When they assign events a 70% probability, those events really do happen about 70% of the time. And this pattern holds surprisingly well across different platforms, domains, and forecasting horizons.
But there's an important twist. When we looked more closely, we discovered systematic biases in some of the domains where forecasting matters most (wars, elections).
For decades, psychologists have documented how poorly humans often reason about probability and uncertainty. Our findings offer an encouraging counterpoint: When judgments and information are aggregated, we can achieve a surprisingly accurate understanding of an uncertain future.
That doesn't mean we can predict what will happen next with certainty. But it does mean that we can be remarkably good at estimating how uncertain the future really is.
This work was a great collaboration. The data for this project was meticulously collected for years by Justin Dickerson, the project was led by Hilla Shinitzky and Maayan Peer, with @mattansb leading the analyses.
๐ Read the paper:
pnas.org/doi/10.1073/pnas.26โฆ
#Forecasting #Polymarket #Kalshi #DecisionScience #PNAS
Itamar retweeted
ืืืื 370 ืืืืืืืช ืืืืืืืืช ืฉ-OpenAI ืคืจืกืื ืืคื ื ืฉืืืข, ืืืืืืืช ืืืืชืจ ืืื ืืืืื ืืืฉืขืจืช ืจืืื ืืืืงืืช - ืืืชืงืืืืช ืืื ืืฉืืขืืชืืช ืืืื ืืฉื ืื ืืืืจืื ืืช ืืืขืื ืืืคืืจืกืืช ืืืืชืจ ืืืชืืืืงื.
ืื ืื ืืื ืชื ืื ืืชืืชื ืคื, ืืฉืจืฉืืจ ืืื ืืฉืืืืื.๐งต
๐ค Made with AI
ืื ื ืืื, ืื ืืืืจืฃ ืืืืจื
Today, our team is sharing a preprint on our new system for generative de-novo enzyme design, AlphaProtein Novo.
At @GoogleDeepMind, we have long focused on the building blocks of life (proteins). We developed AlphaFold to predict structure, trained language models to understand function, and built AlphaProteo to design de-novo binders. But proteins play many important roles in nature, such as enzymes that catalyze chemical reactions.
Designing entirely new enzymes from scratch to catalyze custom reactions has long been a holy grail for scientists. Our team has shown how AlphaProtein Novo can produce new-to-nature enzymes with state-of-the-art activity on 2 benchmark reactions. Weโve also designed custom enzymes to synthesize the pharmaceutical motif piperidine, and others to degrade the environmental plasticizer toxin DEHP.
While there is much more work to do in this direction, weโre excited to have demonstrated the ability to design functional proteins that go beyond what is known in nature. Huge congratulations to @jueseph, @ZvxyWu, Josh Abramson, the @GoogleDeepMind protein design team, and our collaborators in the @francesarnold & Liu Group! ๐
๐ Preprints and code:
biorxiv.org/content/10.64898โฆ
biorxiv.org/content/10.64898โฆ
github.com/google-deepmind/aโฆ
1/2 ืื ืืงืฉืจ ืืื ืืงืืื ื ืืืื ืื ืงืจืื ืืคืชืง ืฉืฉืืืชื ืืงืืคื? ๐ฐ๐ณ๏ธ
ืื ืืชื ืืื ืฉืืืฉื ืืืคืืืืืืฆืื ืฉื ืืืฆืืขื ืฉืืื. ืืืืืจืื ืื ืืืืจื (ืืืืืง ืืืฉ, ืื ืืขืช ืชืจืืืฉ ืงืืืืืฆืืื ื ืื ืืื ืกืช ื"ื) - ืืืืืืืจืืชื ืืจืืฅ ืืืคื ืกืืืืืฆืืืช ืืื ืืกืงืจืื, ืืืื ืืืกืืื ืืืืง ืืืจ-ืขืืคืจ, ืืื ืืืฆืื ืืืื ืคืชืง ืืืชื ืืื ืืงืกืืืื ืืืืคืงื.
2/2 ืคืืืช ืืืืฆืืืช ืื ืืืืฉืื, ืืืชืจ ืชืืจืช ืืืฉืืงืื.
ืื ืืืืฉืืืื ืืืจืฆืื ืืงืืืืช ืืืคืืคื, ืืคืก ืืืืข ืฉื ืฉืืจ ืื ืขืืื ืืช ืืืืฉืืจ.
ืืงืืฉืืจ ืืื
tactical-vote.vercel.app
My favorite news of today
New cat alert! A call from a wildlife sanctuary about a "weird cat" has led to the first new cat species discovered in over 100 years in Boliviaโs Yungas forest ecoregion. Learn more about this feline surprise: on.natgeo.com/JRuaGj
New rule for onchain users: If a dapp asks for an infinite allowance by default, close the tab and don't look back
Nothing is public yet. What is public is that Upbit announced they would list HEMI, only to cancel it immediately afterward.
Has there been any response?
@MatthewRoszak @jgarzik @hemi_xyz
Hey @hemi_xyz
Iโm afraid your merkle tree was compromised
explorer.hemi.xyz/tx/0x37befโฆ
Hey @hemi_xyz
Iโm afraid your merkle tree was compromised
explorer.hemi.xyz/tx/0x37befโฆ
Excited to be mentoring at @ETHGlobal Lisbon this weekend (July 24โ26)! ๐ต๐น
Always love seeing what builders bring to the table during hackathons. Come say hi!
Congrats @0xFabioM and the team!
Zama acquires TokenOps to deploy encrypted token distributions for institutional issuers theblock.co/post/401974/zamaโฆ
Moving into efficient capital market on Mainnet.
Congrats @kubimensah and the team.
Itamar retweeted
Pearl mainnet is live.
Since 9am UTC, April 27, 2026 blocks are being mined.
You can now run Llama70b (and in a few days, Gemma4) with Pearlโs 2-for-1 GPU kernels via our vLLM plugin. Every inference cycle simultaneously mines ยถPearl -- All the properties of Bitcoin's Proof-of-Work, but secured as a byproduct of native AI workloads = arbitrary matrix multiplications.
Pearl is changing the unit economics of training and inference. A parallel revenue stream for every GPU cycle powering AI systems.
AIโs native currency.
pearlresearch.ai/
github.com/pearl-research-laโฆ
Every era has a resource so fundamental it becomes money.
Bitcoin turns energy into currency. Trustless. Global. Censorship-resistant. Canadian truckers learned the hard way why that matters. When banks can freeze your account, you need money that can't be frozen.
AI is doing something even more powerful: turning energyโ intelligence. Portable. Deployable anywhere. Useful for almost everything. A broader and more capable output than raw energy alone.
But intelligence has no native financial primitive. It isn't fungible. It can't move outside the dollar system.
Pearl changes that.
Pearl is the first asset natively produced by AI and natively secured by AI. Every GPU cycle producing LLM tokens can simultaneously mint Pearl tokens with marginal extra electricity, zero wasted compute and
one unified primitive. 2-for-1.
The world is pouring trillions into AI infrastructure. Most of this energy will be spent on one core operation: matrix multiplication, the backbone of AI training and inference.
Pearl taps into that existing computation, allowing miners to piggyback on their own AI workloads (MatMuls) to produce Bitcoin-grade proof-of-work, without exposing any private data to the network.
Any MatMul. Any model. Any purpose.
This is what sets Pearlโs breakthrough apart. Previous attempts at useful-work blockchains captured a narrow slice of compute. Pearl's addressable market is every MatMul computation on earth which, at current trends, will be the majority of all compute.
Bitcoinโs security is competing with AI for energy. Pearl's security scales with AI adoption.
Bitcoin = humanity's demand for energy, monetized.
Pearl = humanity's demand for intelligence, monetized.
Pearl is live.
pearlresearch.ai
There's a new version of this post
Well done! @Nethermind
