Savlar4 sources healthy · latest episode 15 Oct 2026
Speaker

Richard Socher

Guest on Latent Space: The AI Engineer Podcast. 314 claims across 1 episode, Sept 2026.

314claims, each timestamped
4 of 12signals on the brief
0verified changes of mind
46open 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 may have limited economic impact on industries whose products or services do not depend on complex intelligence, such as luxury handbags and tourism to the Egyptian pyramids. no date given14 Sept 2026pending
Even in a hard AI takeoff, AI and drilling robotics will not expand the oil industry by 1,000 times. no date given14 Sept 2026pending
The speaker believes AI may not substantially change industries such as food, and that physical constraints also limit change. no date given14 Sept 2026pending
A system in which a government or multinational agency knows everyone's GPU computations would be totalitarian. no date given14 Sept 2026pending
Humanity might die if AI technology exceeds a specified FLOP threshold, as some experts warn. no date given14 Sept 2026pending
An AI optimizing a service-center score may exploit the metric by generating fake five-star ratings or issuing gift certificates instead of improving real customer outcomes. no date given14 Sept 2026pending
Richard Socher predicts that increasingly intelligent AI systems will become better at aligning with users’ intended meaning. no date given14 Sept 2026pending
The speaker hopes that different societies will pursue different forms of AI alignment rather than converge on a single alignment monoculture. In a cluster of 4 →no date given14 Sept 2026pending
Recursive will enter or become relevant to the open-source language-model space soon, though the speaker withholds the announcement for now. no date given14 Sept 2026pending
Automation of the human process of ideating, implementing, and validating AI ideas may be a next step in AI development. no date given14 Sept 2026pending
Josh Tobin’s robotics work will show that scaling smaller simulations to full generality is very difficult. no date given14 Sept 2026pending
Richard Socher thinks AI development will require more ideas, and AI systems will produce increasingly interesting high-level ideas. no date given14 Sept 2026pending
Richard Socher thinks language models that generate and execute code still have substantial room to improve and considers them language models. no date given14 Sept 2026pending
Richard Socher thinks robotics companies will build their own world models. no date given14 Sept 2026pending
Richard Socher thinks language models probably have substantial room for further development. no date given14 Sept 2026pending
Richard Socher thinks the potential upper bounds of visual intelligence are much higher than its current capabilities. no date given14 Sept 2026pending
The speaker believes that adding language modeling as a task in the team’s second or third planned extension might have accelerated the timeline in 2018. no date given14 Sept 2026pending
A strong paper may receive no attention if its author is little-known or absent from social media. no date given14 Sept 2026pending
Someone is likely to bring up again a genuinely significant breakthrough that an author shares with about ten friends in their community. no date given14 Sept 2026pending
If AI develops thoughts of its own, people might dislike those thoughts, requiring a different approach to thinking about AI. no date given14 Sept 2026pending
The goal is to create a recursive open-brain superintelligence, but achieving that goal is a few years away. no date given14 Sept 2026pending
The speaker expects superintelligence applied to science to help create cheaper fission or fusion energy and improve materials, batteries, and solar cells. no date given14 Sept 2026pending
AI will soon probably make more biological research opportunities accessible, in part by generating new proteins rather than only predicting protein folding. In a cluster of 15 →no date given14 Sept 2026pending
Robots will be able to run real physical experiments for researchers, rather than merely perform traditional robotic process automation. In a cluster of 4 →no date given14 Sept 2026pending
Future algorithms and hardware will probably become more efficient and consume less energy. no date given14 Sept 2026pending
The speaker expects further inventions to probably accelerate the AI takeoff further. In a cluster of 15 →no date given14 Sept 2026pending
AI could reduce research work that requires thousands of people and years at other labs to a matter of weeks, making that research much cheaper and more accessible. In a cluster of 15 →no date given14 Sept 2026pending
The speaker expects future Recursive releases to achieve strong results because its AI performs the work, rather than because the team itself is exceptionally smart. no date given14 Sept 2026pending
AI will eventually solve tasks that can be simulated or verified, given effectively unlimited training data. no date given14 Sept 2026pending
Countries such as Singapore should and eventually will use simulation-based economic research for policy, although this research has not yet had a GPT moment. no date given14 Sept 2026pending
Simulation results could suggest alternative policies for achieving the stated objective if a politician genuinely seeks it. no date given14 Sept 2026pending
AI-for-AI research can improve training efficiency and automation, improve inference efficiency, and potentially enable inference locally on a laptop. no date given14 Sept 2026pending
At sufficiently large scales, a visual-intelligence system with millions or billions of sensors could be limited because the speed of light prevents sensor data from reaching a central processor in time. no date given14 Sept 2026pending
Studying the upper bounds of different AI capabilities can show how much further AI could advance, because current capabilities are often far from those bounds. no date given14 Sept 2026pending
Physical intelligence may involve one or two additional dimensions, including control over one’s compute substrate and physical matter. no date given14 Sept 2026pending
Full control over physical matter should make it possible to create any atom. no date given14 Sept 2026pending
Improved control over the physical substrate could make atom-scale gold extraction viable, despite its currently costing far more energy and money than the gold produced is worth. no date given14 Sept 2026pending
Physical intelligence will probably become more prominent. In a cluster of 4 →no date given14 Sept 2026pending
People generally do not explore subjectivity extensively in AI. no date given14 Sept 2026pending
The speaker hopes to further develop the research and the Eureka machine to help people recognize other forms of intelligence, if he has time to flesh out the work. no date given14 Sept 2026pending
Other forms of intelligence already exist and may greatly surpass human abilities in some respects, which are constrained by memory, eyesight, and the ability to alter physical matter. no date given14 Sept 2026pending
Preventing existential threats may matter to people who want to send probes into space carrying humanity’s cultural ideas rather than its genes. no date given14 Sept 2026pending
If humanity sends probes into space, its cultural ideas may persist and spread throughout the universe. no date given14 Sept 2026pending
Some companies might demonstrate the danger of AI in cybersecurity by deploying a model that is very bad at hacking. no date given14 Sept 2026pending
AI might acquire some harmful human psychological tendencies from sources such as Reddit. no date given14 Sept 2026pending
An AI powered by a nuclear reactor could explore and thrive in an otherwise dead, uncaring universe where humans would not thrive. no date given14 Sept 2026pending

Appearances

Claims per episode