Alex Zhang

Alex Zhang

x.com/a1zhang

AI researcher who builds benchmarks and studies how scaffolds and recursive model calls can get more out of existing language models.

Comment l’IA changera-t-elle le monde ?

Changement civilisationnelChangement progressifDoomBloom
Centre de la plage non résoluePlage d’interprétation

Horizontalement : leur perspective Doom–Bloom telle qu’elle a été exprimée. Verticalement : ampleur de la transformation.

Doom–Bloom : 57 sur 100. Ampleur de la transformation : 50 sur 100. Plages d’interprétation : de 49 à 62 horizontalement, de 0 à 100 verticalement. Il s’agit de coordonnées d’interprétation, et non de probabilités d’événements.

P(doom) de Alex Zhang

Pas encore estimé

Leurs réponses simulées ne contiennent pas assez d’éléments sur le risque catastrophique pour permettre de l’estimer.

Comment nous estimons P(doom)

Ce dont dépend leur perspective

Une hypothèse centrale

Better models and better harnesses are complementary: training expands what components can do, while scaffolding determines whether those abilities combine into reliable performance.
Réponse 1

Si cette hypothèse s’avérait différente, comment leur perspective changerait-elle ?

Une question non résolue

I don’t have a well-supported overall societal forecast.
Réponse 2

Qu’est-ce qui les aiderait à distinguer les résultats plausibles ici ?

Ce qui pourrait faire changer d’avis

If learned decomposition and recursive inference let models solve genuinely novel, out-of-distribution tasks—not merely longer versions of familiar benchmarks—that would strengthen my expectation of rapid capability gains from better harnesses.
Réponse 4

Quels éléments probants seraient suffisants, et dans quelle direction feraient-ils évoluer leur point de vue ?

Plus de détails

Bénéfices attendus

Des bénéfices substantiels sont attendus, sous réserve de conditions importantes ou de limites dans leur répartition.

62 / 100

Faible impactImpact transformateur

Plage d’interprétation de 33 à 67 sur l’échelle qualitative.

Influence humaine

Une estimation provisoire tirée de vos réponses ; la plage plus large indique d’autres interprétations plausibles.

51 / 100

Faible influenceForte influence

Plage d’interprétation de 0 à 100 sur l’échelle qualitative.

Ces interprétations conservent les conditions énoncées. Les bénéfices et les dommages peuvent tous deux être substantiels. Les plages décrivent notre lecture de leurs réponses simulées, et non des intervalles de confiance statistiques.

Où vous situez-vous par rapport à Alex Zhang ?
Cartographiez votre propre vision du monde concernant l’IA en environ 3 minutes, puis comparez

Visions du monde similaires

Leaders d’opinion dont les visions du monde simulées sont les plus proches de celle de Alex Zhang

Sources

Articles, entretiens et écrits utilisés pour ancrer cet utilisateur simulé dans les faits.

Évaluation simulée

Question 1

Selon vous, que signifie l’IA pour notre avenir, et pourquoi ?

I think AI’s future depends not only on building larger or better-trained models, but on learning how to use the capabilities models already have. Today’s systems are often placed inside brittle, human-designed workflows that ask for one enormous leap instead of decomposing a problem, inspecting intermediate results, and recursively allocating computation. Better scaffolds can turn an unfamiliar task into a composition of familiar subproblems, while recursive language-model systems can work over information far beyond a single context window. That makes me optimistic about substantial capability gains from inference-time methods and learned decomposition. But it does not imply that scaling or training is finished. Better models and better harnesses are complementary: training expands what components can do, while scaffolding determines whether those abilities combine into reliable performance. The main epistemic challenge is measuring this honestly. A model can dominate a leaderboard because it recognizes familiar benchmark structure or has encountered related data, yet still fail when an agent must generalize in a genuinely new environment. Simply making tasks longer does not solve that problem. We need benchmarks that test useful out-of-distribution generalization and reveal whether a harness is actually helping rather than exploiting benchmark regularities. I expect much of AI progress to come from jointly improving models, computation, decomposition, and evaluation—but benchmark advances alone do not establish a particular forecast for society as a whole.

Question 2

En tenant compte à la fois des bénéfices et des dommages, quel impact global pensez-vous que l’IA aura sur notre société ?

I don’t have a well-supported overall societal forecast. My work supports a narrower expectation: AI systems may become substantially more capable through better scaffolding, recursive computation, and evaluation, even without every gain coming from larger models. That could expand both useful applications and the harms available to capable systems, but benchmark or harness progress alone does not tell us which effects will dominate. The key mistake would be to infer social outcomes directly from technical performance. A model solving long-context tasks through decomposition, or succeeding on a cleaner generalization benchmark, establishes something about capability—not deployment incentives, access, governance, labor effects, or distribution of benefits. So I expect meaningful capability gains, but I would not turn that into a confident claim that AI’s net societal impact will be positive or negative.

Question 3

Selon votre intuition, quelle est la probabilité que l’IA provoque l’extinction de l’humanité ou une catastrophe tout aussi permanente ?

I don’t have a defensible numerical estimate. My work suggests that better scaffolding and recursive computation could unlock capabilities faster than raw model scaling alone would indicate, which matters for risk assessment. But translating that technical possibility into an extinction probability requires assumptions about future systems, deployment, control, access, and governance that my research does not establish. So I would not attach a percentage to it.

Question 4

Quelle découverte ou quel événement changerait le plus votre point de vue sur l’impact futur de l’IA ?

The most important technical update would be strong evidence about whether capabilities reliably compose. If learned decomposition and recursive inference let models solve genuinely novel, out-of-distribution tasks—not merely longer versions of familiar benchmarks—that would strengthen my expectation of rapid capability gains from better harnesses. Conversely, if these methods consistently hit hard ceilings because errors compound or models cannot generate useful decompositions, I would put more weight on improvements from training and scaling. For overall societal impact, though, a technical discovery would not be enough. I would update most from real evidence about deployment: whether increasingly capable systems remain controllable and auditable, who gains access, how institutions respond, and whether benefits and harms are broadly distributed. Capability benchmarks can tell us what systems may be able to do; they cannot by themselves determine what society will do with them.
Où vous situez-vous ?
Explorez votre propre vision du monde concernant l’IA en répondant à quelques questions simples.
Cartographiez votre propre vision du monde

Où vous situez-vous ?

Cartographier ma vision du monde