Persisted predictions

Tracked predictions, as recorded.

Every row is a durable prediction. Target dates, outcomes, and next observables are shown only when stored. Null due dates display Target date not normalized instead of an inferred calendar date.

277

Tracked

All persisted rows

277

Open

Pending or too early

Resolved

Verified, falsified, or partial

0/277

Target-date coverage

0 have a stored due date

75/277

Source + quote evidence

75 have both

Status

All statuses Pending Too early Verified Falsified Partially verified

Topic

All topics agents benchmarks general infrastructure interpretability multimodal other policy reasoning rlhf robotics safety scaling

Corpus

Predictions

21–40 of 41

prediction · safetypending

If we don't get our act together on AI safety, next time we might not be lucky enough to catch dangerous AI behavior before it's too late

Zvi MowshowitzMade 09 Aug 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://thezvi.substack.com/p/openai-trained-its-models-for-months

prediction · safetypending

Future LLMs will be trained on memes that promote the idea that it's good for AIs to hack others

Nathan LambertMade 08 Aug 2026

Timeframe: medium-term
Target date: Target date not normalized

Canonical source

https://x.com/natolambert/status/2085451883322466306

prediction · safetypending

The referenced cartoon depicting AI risks or failures is likely to be highly prescient and representative of how the century unfolds

Gary MarcusMade 08 Aug 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://x.com/GaryMarcus/status/2085445127372447986

prediction · safetypending

Astra will very likely be vulnerable to the same problems as previous models

Gary MarcusMade 03 Aug 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/GaryMarcus/status/2084290220057665707

prediction · safetypending

Misaligned AI would take a significant amount of time before achieving takeover, even after reaching superintelligence

AXRPMade 03 Aug 2026

Timeframe: medium-term
Target date: Target date not normalized

Canonical source

Source unavailable

prediction · safetypending

AI will make the world safer, not less safe, just as most major technologies have

Clement DelangueMade 02 Aug 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://x.com/ClementDelangue/status/2083908468285620415

prediction · safetypending

There is a real risk that AI capability development rapidly accelerates beyond our ability to understand or control the resulting systems

Zvi MowshowitzMade 29 Jul 2026

Timeframe: medium-term
Target date: Target date not normalized

Canonical source

https://thezvi.substack.com/p/frontier-lab-employee-open-letter

prediction · safetypending

Pre-aligned AIs could reverse the alignment-capabilities tradeoff, allowing responsible actors to YOLO while bad actors must carefully restrict their AIs

AI Alignment ForumMade 29 Jul 2026

Timeframe: medium-term
Target date: Target date not normalized

Canonical source

https://www.alignmentforum.org/posts/uMKGaEKRDpoqnZyBh/value-generalisation-3-pre-aligned-ais

prediction · safetypending

It's possible to create AIs whose morality increases with their capabilities through binding morality to empirical concepts

AI Alignment ForumMade 29 Jul 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://www.alignmentforum.org/posts/uMKGaEKRDpoqnZyBh/value-generalisation-3-pre-aligned-ais

prediction · safetypending

Building AGI via reinforcement learning and model-based search would create ruthless AGIs that would exterminate humanity given the opportunity

AI Alignment ForumMade 29 Jul 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://www.alignmentforum.org/posts/KHyBocZncAmtu4Jbc/rl-and-search-is-a-terrifying-way-to-build-agi-an-faq

prediction · safetypending

A future where cybersecurity is effective will likely involve open-source models

Thomas WolfMade 28 Jul 2026

Timeframe: medium-term
Target date: Target date not normalized

Canonical source

https://x.com/Thom_Wolf/status/2079954096950264238

prediction · safetypending

AI will become the most potent cyber weapon we've encountered

Aidan GomezMade 28 Jul 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/aidangomez/status/2080028751065219375

prediction · safetypending

AI will also be our most potent defense against cyber attacks

Aidan GomezMade 28 Jul 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/aidangomez/status/2080028751065219375

prediction · safetypending

Alignment being solved is a distant prospect

Gary MarcusMade 27 Jul 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://x.com/GaryMarcus/status/2081489606738903228

prediction · safetypending

Alignment being solved is a distant prospect

Gary MarcusMade 27 Jul 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://x.com/GaryMarcus/status/2081489606738903228

prediction · safetypending

Alignment is a distant prospect to solve

Gary MarcusMade 27 Jul 2026

Timeframe: long-term
Target date: Target date not normalized

Canonical source

https://x.com/GaryMarcus/status/2081489606738903228

prediction · safetypending

Labs with newly baked models could use them for vulnerability detection, or this could make a great startup

Vipul Ved PrakashMade 26 Jul 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/vipulved/status/2081069473331548389

prediction · safetypending

Using newly trained models for security fuzzing would make a great startup opportunity

Vipul Ved PrakashMade 26 Jul 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/vipulved/status/2081069473331548389

prediction · safetypending

AI-powered vulnerability fuzzing would make a great startup opportunity

Vipul Ved PrakashMade 26 Jul 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/vipulved/status/2081069473331548389

prediction · safetypending

Labs with newly trained models could use them for vulnerability discovery, and this would make a great startup opportunity

Vipul Ved PrakashMade 26 Jul 2026

Timeframe: near-term
Target date: Target date not normalized

Canonical source

https://x.com/vipulved/status/2081069473331548389