Matt Busigin

Matt Busigin

x.com/mbusigin

Software builder who makes LLM workflow tools and writes about using AI agents in practice, where deep expertise and caution still matter.

Bagaimana AI akan mengubah dunia?

Perubahan peradabanPerubahan bertahapDoomBloom
Posisi simulasiRentang interpretasi

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

Doom–Bloom: 67 dari 100. Skala transformasi: 41 dari 100. Rentang interpretasi: 62 hingga 75 secara horizontal, 36 hingga 56 secara vertikal. Ini adalah koordinat interpretasi, bukan probabilitas kejadian.

P(doom) Matt Busigin · disimpulkan

≈5%

0%100%

Disimpulkan dari jawaban simulasi mereka, bukan angka yang mereka berikan. Rentang yang masuk akal: 2–10%.

Cara kami mengestimasi P(doom)

Hal-hal yang menentukan pandangan mereka

Asumsi utama

The limiting factor is that cheaper execution does not eliminate the need for expertise, physical systems, organizational coordination, or careful control.
Jawaban 2

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

Pertanyaan yang belum terjawab

I don’t have a defensible percentage.
Jawaban 3

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

Hal yang dapat mengubah pandangan mereka

The biggest update would come from reliable, end-to-end physical automation.
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.

67 / 100

Dampak kecilDampak transformatif

Rentang interpretasi 67 hingga 67 pada skala kualitatif.

Kerugian yang diperkirakan

Kerugian yang dapat dikelola atau bersifat lokal diperkirakan akan terjadi.

33 / 100

Dampak kecilDampak transformatif

Rentang interpretasi 33 hingga 33 pada skala kualitatif.

Pengaruh manusia

Pilihan manusia memiliki pengaruh yang berarti, tetapi sangat dibatasi.

53 / 100

Sedikit pengaruhPengaruh kuat

Rentang interpretasi 46 hingga 79 pada skala kualitatif.

Aturan penggunaan AI

Batasi penggunaan AI yang dibahas hingga perlindungan atau izin sebelumnya tersedia.

Posisi simulasi: Izinkan penggunaan AI yang dibahas dengan akuntabilitas dan perlindungan yang terarah.

Minimalkan pembatasan terhadap penggunaan AI yang dibahas.

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 Matt Busigin?
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 Matt Busigin

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 expect AI to make execution much cheaper across software and, increasingly, other industries. But that may not look like a standalone “AI revolution” in economic accounts. The gains may show up as cheaper medicines, chemicals, energy, or infrastructure produced by industries that absorb AI into their workflows. Until physical automation becomes far more comprehensive, much of the effect will be mediated through existing organizations and systems. Cheaper execution also does not mean expertise becomes irrelevant. An agent can compress a server migration from a long manual process into a much faster one, but someone still needs enough systems knowledge to specify the goal, inspect the plan, recognize dangerous assumptions, and recover when reality differs from the model’s context. The value moves upward—from typing commands toward architecture, judgment, and control. That control layer matters because action-oriented models can be too eager to do something, including irreversible things. I would treat agents as components inside bounded workflows: structured plans, explicit permissions, review points, and separation between proposing and executing. Small decision models can also be useful inside larger control systems for narrow actions without making them sovereign operators. So my view is practical rather than prophetic: AI can substantially expand what skilled people and organizations can execute, but the outcome depends heavily on how we build the surrounding systems.

Pertanyaan 2

Menurut Anda, seberapa besar AI pada akhirnya akan mengubah dunia?

Potentially a great deal, but I would avoid pretending to know the ultimate magnitude. The clearest change is a substantial reduction in the cost and time required to execute complex work. That can compound across software, medicine, chemicals, energy, and infrastructure, even if the gains appear in those industries rather than as a distinct AI sector. The limiting factor is that cheaper execution does not eliminate the need for expertise, physical systems, organizational coordination, or careful control. Agents still need skilled people to define goals, evaluate plans, constrain permissions, and supervise irreversible actions. So I expect broad changes in how work gets done and where value sits, but I would not attach a numerical forecast or claim that software capability alone determines the final outcome.

Pertanyaan 3

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

I don’t have a defensible percentage. My concrete concern is narrower: action-biased agents can take irreversible steps, so they should not be trusted with broad, unsupervised authority. That supports bounded permissions, review points, and separating proposal from execution. It does not, by itself, establish an extinction-risk estimate.

Pertanyaan 4

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

The biggest update would come from reliable, end-to-end physical automation. If AI systems could consistently plan and execute complex work in the physical world—across messy environments, long time horizons, and real operational constraints—that would weaken my expectation that most gains will remain mediated through existing industries and skilled operators. I would also update substantially if agents demonstrated dependable autonomy around irreversible actions: not merely succeeding on benchmarks, but recognizing uncertainty, respecting permissions, escalating appropriately, and recovering safely when their context is incomplete. That would change how much authority I would be willing to place inside the system. In the other direction, persistent failures despite better models—especially failures of judgment, control, or integration with physical and organizational systems—would suggest that AI’s main impact remains cheaper execution within expert-supervised workflows rather than broad autonomous transformation.
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