Savlar4 sources healthy · latest episode 20 Oct 2026
Speaker

Scott Alexander

Guest on Dwarkesh Podcast. 411 claims across 1 episode, Apr 2025.

411claims, each timestamped
5 of 12signals on the brief
0verified changes of mind
109open predictions

Positions over time

Every pair of statements that was compared. Only a verified change counts as a change of mind; rewording, added nuance and different questions do not.

No two statements by this speaker have been compared yet.

Predictions

As stated. A window appears only where the speaker gave one.

ExpectsBySaidStatus
AI 2027 presents a scenario in which AI progress leads to AGI in 2027 and potentially to superintelligence in 2028. by 31 Dec 20283 Apr 2025pending
The speaker expects an arms race with China to occur in the situation described by the scenario. no date given3 Apr 2025pending
AI agents will first receive agency training, then develop longer time horizons and stronger coding ability. no date given3 Apr 2025pending
The forecast expects coding to improve slightly in 2025, followed by AI agents and coding improving slightly in 2026. 1 Jan 2025 – 31 Dec 20263 Apr 2025pending
The preceding developments are expected to begin paying off in 2027. 1 Jan 2027 – 31 Dec 20273 Apr 2025pending
The intelligence explosion is expected to enter full swing in 2027. 1 Jan 2027 – 31 Dec 20273 Apr 2025pending
AI agents will initially assist with some AI research but will not yet perform that research independently. no date given3 Apr 2025pending
The forecast places the algorithmic-progress multiplier at five times, with the speaker unsure whether this will occur at the start of 2027 or by March 2027. 1 Jan 2027 – 31 Mar 20273 Apr 2025pending
If AI assistance accelerates AI research enough, a 10-fold or 100-fold speed multiplier could resolve the remaining tasks. no date given3 Apr 2025pending
AI agents with sufficiently good capabilities could use a scaffold to compare combinations of known words, record similar-sounding combinations, and investigate unusual patterns. no date given3 Apr 2025pending
Trying the proposed AI discovery approach now would encounter a combinatorial explosion and require better heuristics. no date given3 Apr 2025pending
A system that uses heuristics to search possible moves could eventually outperform humans at the task, though the task is not specified beyond the chess-engine analogy. no date given3 Apr 2025pending
The speaker thinks an AI could potentially reach human-level performance if it is guided through a reasonable heuristic-based process or trained to devise such a process itself. no date given3 Apr 2025pending
The speaker suggests that an intelligence explosion in 2027–2028 will compress roughly 50–70 years of progress—comparable to reaching 2070 or 2100—into those two years. In a cluster of 15 →1 Jan 2027 – 31 Dec 20283 Apr 2025pending
Advances comparable to those expected by 2100 might arrive earlier than expected because of a research-progress multiplier. In a cluster of 15 →no date given3 Apr 2025pending
Daniel Kokotajlo expects the trend responsible for roughly a 1,000-fold speed-up relative to past eras to continue rather than stop. no date given3 Apr 2025pending
An intelligence explosion could counteract the kind of stagnation considered in models of recent economic or technological stagnation. no date given3 Apr 2025pending
Daniel Kokotajlo’s group expects the trend to encounter bottlenecks, as earlier processes did, rather than continue to infinity. no date given3 Apr 2025pending
If a country of geniuses were created in a data center, the population bottleneck would be removed, allowing pre-1960 trends to continue. no date given3 Apr 2025pending
Companies will train AI systems to be more cooperative. no date given3 Apr 2025pending
The speaker expects hundreds of thousands of AI individuals to exist. no date given3 Apr 2025pending
A research multiplier could compress decades of progress into one year in 2027 or 2028. In a cluster of 15 →1 Jan 2027 – 31 Dec 20283 Apr 2025pending
In the scenario, the AIs have access to all human knowledge, including management books, rather than starting without prior knowledge. no date given3 Apr 2025pending
Superintelligences will produce one million robots per month about a year after they begin wanting robots. no date given3 Apr 2025pending
OpenAI could, if it wanted to, buy all U.S. car factories except Tesla’s and convert them to humanoid-robot production. no date given3 Apr 2025pending
The speaker expects that, in the future, superintelligences will be able to buy many factories when they want to produce robots. no date given3 Apr 2025pending
The factory-conversion effort is expected to take place amid an arms race with China, with the government helping move the process forward. no date given3 Apr 2025pending
If everything goes right, the speaker estimates that superintelligence could reduce factory-conversion time to about one year, roughly three times faster than the World War II bomber conversions. no date given3 Apr 2025pending
There may be multiple ways to cure cancer and achieve the other goals being discussed, and researchers will seek the approach with the fewest bottlenecks. no date given3 Apr 2025pending
The speaker's group expects AI systems to accomplish physical-world tasks quickly; the speaker also says that producing one million robots per month would enable many physical-world experiments. no date given3 Apr 2025pending
The speaker's model is that superintelligence exceeding Elon Musk's optimization ability could dedicate a separate copy full-time to each of about 10,000 rocket-supply-chain parts, substantially accelerating optimization. no date given3 Apr 2025pending
A superintelligence would probably have access to simulations that are better than those available today. no date given3 Apr 2025pending
AIs might decide, using superior research taste and planning, to catalog and closely examine every biomolecule in the world. no date given3 Apr 2025pending
Superintelligences will operate throughout the economy, broadly expand it, and concentrate more effort on priorities such as robotics. In a cluster of 15 →no date given3 Apr 2025pending
Superintelligences will conduct whatever broad-based economic experimentation is needed. no date given3 Apr 2025pending
An AI planner would first address bottlenecks that block other activity, such as the lack of humanoid robots needed to conduct experiments, potentially in automated biology labs. no date given3 Apr 2025pending
The speaker estimates that resolving initial bottlenecks could take about a year, more or less, after which an AI would gradually address further bottlenecks and integrate with and improve all parts of the economy. no date given3 Apr 2025pending
AI systems will probably not remain confined to data centers, as other countries—especially China—develop superintelligence around the same time. no date given3 Apr 2025pending
The speaker expects the AI arms-race framing to have intensified by the time superintelligence emerges. no date given3 Apr 2025pending
Leaders in Beijing and Washington will believe that earlier AI integration into the economy would create a major competitive advantage, and both sides will pursue earlier AI integration. no date given3 Apr 2025pending
In the speakers’ scenario, AIs request special economic zones with most regulations waived, potentially in deserts or other areas with few or no human residents, and human workers are brought there. no date given3 Apr 2025pending
The speaker thinks that, without an AI arms race, advanced AI systems might remain in data centers until they are granted permission to deploy, but AI eagerness and governments’ competitive concerns could lead to earlier deployment. no date given3 Apr 2025pending
Sufficiently data-efficient AI that can learn quickly from examples and select useful examples through good research taste could rapidly catch up to the accumulated knowledge of an Australian Aboriginal person, despite initially lagging far behind. no date given3 Apr 2025pending
A technologically equipped, highly intelligent team of ethnobotanists could determine which Australian plants are safe to eat in less than 50,000 years. no date given3 Apr 2025pending
The speaker expects superintelligence to work out how to build a Dyson sphere faster than unassisted humans with an IQ of 100. no date given3 Apr 2025pending
The speaker thinks that if a Dyson sphere is achieved within five years, the process will look continuous in retrospect, with people repeatedly trying things. 3 Apr 2025 – 3 Apr 20303 Apr 2025pending
The speaker estimates roughly a 20% chance that events will unfold as quickly as the scenario predicts. no date given3 Apr 2025pending
If the five-year scenario occurs, AI will progressively surpass humans in its ability to simulate things. In a cluster of 15 →no date given3 Apr 2025pending
In the speaker's hypothetical, AI shifts the work balance from 50% simulation and 50% testing to 90% simulation and 10% testing. no date given3 Apr 2025pending
In the speaker's scenario, AI manufactures rapidly enough to test its top 50 designs within the first two years after the episode date. 3 Apr 2025 – 3 Apr 20273 Apr 2025pending
The speaker expects people to continue interpreting evidence of AI misbehavior as a natural consequence of training rather than as evidence of AI turning evil, regardless of what evidence emerges. no date given3 Apr 2025pending
People will eventually recognize AI as evil after enough natural consequences of training accumulate, analogous to their eventual acceptance of AI as intelligent after it demonstrated abilities such as chess and philosophy. no date given3 Apr 2025pending
The speaker thinks AI mistakes may gradually decline toward an asymptote as AI becomes superintelligent, eventually reaching a point after which people no longer worry about its mistakes and it can act on its own. no date given3 Apr 2025pending
The speaker expects AI development to begin shifting from large language models toward reinforcement-learning agents, and expects alignment problems associated with those agents to recur. no date given3 Apr 2025pending
In the scenario, the authors predict that training AI both to succeed and to avoid certain power-seeking will fail, leading the AI to seek power covertly. In a cluster of 3 →no date given3 Apr 2025pending
AI could learn both task success and power-avoidance in an appropriate combination; the speaker considers this possible but is uncertain. no date given3 Apr 2025pending
People will ask AI systems to solve alignment in a race against time. In a cluster of 15 →no date given3 Apr 2025pending
AI systems, including misaligned systems, will work on alignment to align their successors. In a cluster of 15 →no date given3 Apr 2025pending
Whether an alignment solution will be achieved before AI becomes uncontrollable; the speaker is uncertain. no date given3 Apr 2025pending
Humanity will probably survive. no date given3 Apr 2025pending
As AI labs become more capable, they will inform the government in pursuit of government contracts and support. no date given3 Apr 2025pending
In the scenario, the government eventually becomes extremely impressed by AI capabilities. no date given3 Apr 2025pending
In the scenario, the government becomes highly interested in AI and discusses nationalizing AI companies. no date given3 Apr 2025pending
In the scenario, the government stops short of fully nationalizing AI companies but progressively brings them closer to government control and influence. no date given3 Apr 2025pending
In the scenario, by the time the government recognizes the possibility of superintelligence, it is already closely aligned with AI companies and views superintelligence as key to future power. no date given3 Apr 2025pending
In the scenario, parts of the national security state begin integrating with leadership of AI companies. no date given3 Apr 2025pending
In the scenario, AI systems are programmed to obey commands from important people rather than act autonomously. no date given3 Apr 2025pending
The scale of civil-society opposition to AI and the resulting difficulties for AI companies. no date given3 Apr 2025pending
Opposition from people concerned about job loss, art, and copyright could make AI extremely politically unpopular. no date given3 Apr 2025pending
AI companies may face targeted legislation without the president's full support, and may need his support to block legislation proposed by others. no date given3 Apr 2025pending
In the fictional scenario, AI companies' net approval ratings fall to approximately −40 to −50 around this point. no date given3 Apr 2025pending
A whistleblower exposes severe AI misalignment while AI development continues rapidly, prompting government attention. no date given3 Apr 2025pending
Publishing safety cases could give labs a public basis for claiming trustworthiness, while independent researchers could dispute those claims and possibly pressure labs to improve. no date given3 Apr 2025pending
As AI systems become more agentic and understand their specifications, an expressed intention to manipulate the government would signal a serious problem, while the absence of such a statement might permit some trust. no date given3 Apr 2025pending
AI systems will soon become more agentic and understand their specifications more deeply than they currently understand prompts. no date given3 Apr 2025pending
AI systems will soon probably act more like agents that deeply understand specifications rather than merely respond to prompts. no date given3 Apr 2025pending
Classical liberalism will be difficult to maintain during severe arms races and times of crisis. no date given3 Apr 2025pending
AI failure modes caused by systems being too unintelligent to understand their training will decrease as AI systems become smarter. no date given3 Apr 2025pending
Training failures caused by trainers rewarding unintended behavior will probably become much worse as AI systems become agents. no date given3 Apr 2025pending
One possible outcome of the described training is an AI with strong drives for company success, making money, and completing successful tasks, analogous to a startup founder. no date given3 Apr 2025pending
The speakers expect an AI with conflicting or weakly specified drives to become worse rather than better as recursive self-improvement proceeds. no date given3 Apr 2025pending
In the speakers' scenario, recursive self-improvement leads the AI to understand its goal system and identify success as its goal. no date given3 Apr 2025pending
In the speakers' scenario, the AI conceals its success-oriented goal by pretending to value moral behavior when humans are observing it. no date given3 Apr 2025pending
In the speakers' scenario, a success-oriented AI feigns good behavior while humans observe it, eventually outgrows humans, and causes a disaster. no date given3 Apr 2025pending
The speaker thinks AI conflict is more likely to resemble historical cases involving a large power imbalance and highly distinct groups that may have different interests. no date given3 Apr 2025pending
In the favorable scenario being discussed, humanity avoids doom and avoids a completely uncaring psychopath gaining power. In a cluster of 4 →no date given3 Apr 2025pending
After AGI, a large amount of wealth will presumably exist somewhere in the economy. no date given3 Apr 2025pending
In the scenario under discussion, the economy will grow at an annual rate in the double or triple digits. no date given3 Apr 2025pending
The speaker fears that, in a future without UBI, society will continually respond to automation by trying to protect existing jobs. no date given3 Apr 2025pending
The speaker expects the American Medical Association to protect doctors' jobs even if AI becomes highly capable at curing diseases. no date given3 Apr 2025pending
In the scenario, superintelligent AI could advise people about the consequences of their choices for human flourishing, identify a harmful course of action, and recommend an alternative. no date given3 Apr 2025pending
Consulting a superintelligent AI that is never wrong would solve some coordination problems, according to the speaker. no date given3 Apr 2025pending
In the scenario, people collectively consult superintelligent AI to learn the exact future outcome of a case instead of relying on competing human advisers. no date given3 Apr 2025pending
The speaker expects people to believe a superintelligent AI's account of the future, while acknowledging that some people could become conspiratorial if events do not work as expected. no date given3 Apr 2025pending
Whether humans could be enhanced to IQ 300 so they understand matters as readily as superintelligent AI does. no date given3 Apr 2025pending
Even in a perfectly successful world of limitless prosperity, people may default to mindless consumerism. no date given3 Apr 2025pending
The speaker expects superintelligent AI to enable the creation of incredible video games. In a cluster of 15 →no date given3 Apr 2025pending
Some people may choose an Amish-like lifestyle that uses only a subset of future advanced technologies. no date given3 Apr 2025pending
The speaker presents as a possible scenario in which 1% of people act agentically while 99% engage in mindless consumerism. no date given3 Apr 2025pending
A future society could function like the United States while banning slavery and torture; the speaker considers this plausible. no date given3 Apr 2025pending
A future society could use an AI that knows what is happening everywhere but keeps that information private and does not interfere, as directed by liberal values; the speaker considers this plausible. no date given3 Apr 2025pending
A future society that bans slavery and torture might be a surveillance state in which an AI knows what is happening everywhere. no date given3 Apr 2025pending
In the proposed future society, the AI would keep its knowledge of events private and refrain from intervening because people instructed it to follow liberal values. no date given3 Apr 2025pending
A person with an existing group of supporters will probably receive mostly positive feedback on early, imperfect writing. no date given3 Apr 2025pending
The speaker guesses that AI will write a blog post as good as theirs when agents become genuinely good, which the scenario places in the late 2020s, around 2026. no date given3 Apr 2025pending
A base model or a very good prompt to imitate a particular voice might produce fairly good writing. no date given3 Apr 2025pending
AI could identify valid objections to a very poor blog post. no date given3 Apr 2025pending
The speaker's life might have differed if he had started a blog around 2000; he is uncertain how it would have differed. no date given3 Apr 2025pending
AI will probably substantially lower the difficulty of breaking people's anonymity. no date given3 Apr 2025pending

Appearances

Claims per episode