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.

Bagaimana AI akan mengubah dunia?

Perubahan peradabanPerubahan bertahapDoomBloom
Titik tengah rentang yang belum ditentukanRentang interpretasi

Mendatar: pandangan Doom–Bloom yang mereka ungkapkan. Ke atas: skala transformasi.

Doom–Bloom: 57 dari 100. Skala transformasi: 50 dari 100. Rentang interpretasi: 49 hingga 62 secara horizontal, 0 hingga 100 secara vertikal. Ini adalah koordinat interpretasi, bukan probabilitas kejadian.

P(doom) Alex Zhang

Belum diestimasi

Jawaban simulasi mereka tidak cukup membahas risiko katastrofik untuk memperkirakannya.

Cara kami mengestimasi P(doom)

Hal-hal yang menentukan pandangan mereka

Asumsi utama

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

Jika asumsi ini ternyata berbeda, bagaimana pandangan mereka akan berubah?

Pertanyaan yang belum terjawab

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

Apa yang akan membantu mereka membedakan hasil-hasil yang masuk akal di sini?

Hal yang dapat mengubah pandangan mereka

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.
Jawaban 4

Bukti apa yang akan memadai, dan ke arah mana bukti itu akan mengubah pandangan mereka?

Detail lebih lanjut

Manfaat yang diperkirakan

Manfaat besar diperkirakan akan terwujud, dengan syarat penting atau keterbatasan distribusi.

62 / 100

Dampak kecilDampak transformatif

Rentang interpretasi 33 hingga 67 pada skala kualitatif.

Pengaruh manusia

Estimasi sementara dari jawaban Anda; rentang yang lebih lebar menunjukkan penafsiran lain yang masuk akal.

51 / 100

Sedikit pengaruhPengaruh kuat

Rentang interpretasi 0 hingga 100 pada skala kualitatif.

Interpretasi ini mempertahankan kondisi yang mereka nyatakan. Manfaat dan kerugian dapat sama-sama besar. Rentang tersebut menggambarkan cara kami membaca jawaban simulasi mereka, bukan interval kepercayaan statistik.

Di mana posisi Anda dibandingkan dengan Alex Zhang?
Petakan pandangan dunia AI Anda sendiri dalam waktu sekitar 3 menit, lalu bandingkan

Pandangan dunia serupa

Pemimpin opini dengan pandangan dunia simulasi yang paling mendekati pandangan Alex Zhang

Sumber

Artikel, wawancara, dan tulisan yang digunakan sebagai landasan bagi pengguna simulasi ini.

Penilaian Simulasi

Pertanyaan 1

Menurut Anda, apa arti AI bagi masa depan kita—dan mengapa?

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.

Pertanyaan 2

Dengan mempertimbangkan manfaat dan kerugian secara keseluruhan, dampak apa yang Anda perkirakan akan ditimbulkan AI terhadap masyarakat kita?

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.

Pertanyaan 3

Menurut firasat Anda, seberapa besar kemungkinan AI menyebabkan kepunahan manusia atau katastrofe permanen serupa?

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.

Pertanyaan 4

Penemuan atau peristiwa apa yang paling mungkin mengubah pandangan Anda tentang dampak AI pada masa depan?

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.
Di mana posisi Anda?
Jelajahi pandangan dunia AI Anda sendiri dengan menjawab beberapa pertanyaan sederhana.
Petakan pandangan dunia Anda sendiri

Di mana posisi Anda?

Petakan pandangan dunia saya