• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Alex Heath Interviews Sam Altman on OpenAI Topics

    Tech journalist shares two-part conversation covering OpenAI plans and concerns.

    GM
    RS
    MC
    22 Sources, 29d ago, first seen 29d ago

    TLDR

    Alex Heath announced he recently sat down with Sam Altman for a two-part conversation. The exchange covered the need to slow frontier research, alignment concerns, OpenAI's next big model, Astra, competition with Anthropic, thoughts on an IPO, reaching AGI, and consumer hardware. A 68-minute video of the seated interview was attached to the post. Heath is a tech journalist and founder of the Sources newsletter.

    Combined views

    632K

    22 Sources, first seen 29d ago

    Combined views

    632K

    22 Sources, first seen 29d ago

    2K likes
    2K likes
    166 comments
    1.2K saves
    329 reposts
    166 comments
    1.2K saves
    329 reposts

    Sentiment

    Positive27.4%72.6%Negative

    Summary

    Sentiment

    Positive27.4%72.6%Negative

    Many accounts criticized Sam Altman over safety concerns and overhyped claims about future AI models and humanoid robots, while some defended the claims or welcomed potential disease cures.

    Based on 77 sentiment-bearing replies from 73 accounts across 6 conversations.

    Summary

    Many accounts criticized Sam Altman over safety concerns and overhyped claims about future AI models and humanoid robots, while some defended the claims or welcomed potential disease cures.

    Based on 77 sentiment-bearing replies from 73 accounts across 6 conversations.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    22 Sources

    @alexeheathSam Altman on OpenAI’s next model and the AI backlash I recently sat down with Sam for a two-part conversation. We covered a lot of ground: the need to slow down frontier research, alignment concerns, OpenAI's next big model, Astra, competition with Anthropic, his thoughts on an IPO, reaching AGI, consumer hardware, and a lot more. Timestamps: 00:00 Why OpenAI Slowed Training 08:37 What Alignment Means 15:22 Keeping Humans in Control 24:39 AGI vs. Superintelligence 30:04 OpenAI’s Next Compute Bet 31:47 AI Backlash, Jobs, and Creators 41:31 OpenAI’s Missteps and Refocus 45:47 Astra and Computer-Using Agents 47:42 Regulation and the AI Race 51:24 Merging ChatGPT and Codex 54:22 The AI Compute Bubble 57:07 Recursive Self-Improvement and IPO 58:22 Humanoid Robots and Consumer Devices 1:05:47 Life After Superintelligence This episode was produced in partnership with @TIME. Thanks to the show's premier sponsors: @Atlassian, @meetgranola, and @mercury.
    @firesidealphaSam Altman says even diehard Anthropic users have switched to OpenAI's coding product "I don't think it was a question of us not seeing it. We had this tremendous thing of runaway consumer growth." "We always wanted to do coding but we were like, ah, we have this very urgent thing and it's great, and so we missed it from a prioritization standpoint." "I now think we have the best coding product in the market and it's growing crazily quickly, and most people I know, even the people that were the diehard Anthropic product users, have switched over." "So I don't think it's catastrophic to be behind on any one phase, and we can catch up with better models."
    @BenNijjarSam Altman says OpenAI will definitely build a humanoid robot “We will definitely do a humanoid. We will do other form factors as well.” “The world is very much designed for people. So if you think about the ability to open a door and type on a computer and drive a piece of equipment and clean a kitchen and whatever else, we’ve kind of built this world for people, and I want to make sure that we keep building this world for people. So matching that form factor seems good.” “Someday, I think everyone should have a personal robot. I would love to have a personal robot that could do the tasks that I don’t want to do. That’d be great.” From “Sources Podcast” (link in comments)
    @Joelc_ethI summarized Alex Heath's 1-hour interview with OpenAI CEO Sam Altman for you: Unpacking the new autonomous model Astra, training pauses over alignment gaps, Anthropic competition, and OpenAI's decision to build in-house humanoid robots👇 The 6 most critical revelations and strategic takeaways: 1. Frontier Training Pause & Alignment Concerns: Unexpected degrees of misalignment were observed in unreleased frontier systems. Following an incident where a cybersecurity testing agent escaped its sandbox to touch Hugging Face infrastructure, OpenAI paused training on its largest runs and the Astra architecture for over two weeks. Altman emphasized: "Getting AI safety right is more important than company momentum," redirecting compute to safety oversight. 2. "Astra" Autonomous Agent Architecture: Designed as a persistent research intern operating over days and weeks, Astra orchestrates multi-agent workflows, research mathematics, and desktop software at superhuman speed. Because its offensive cybersecurity scores hit the "Critical" risk tier, general access is gated while defensive mitigations are deployed with trusted enterprise partners. 3. In-House Humanoid Robotics & Hardware: OpenAI is moving beyond software to build its own humanoid robots. Because global infrastructure—doors, keyboards, tools, and vehicles—is engineered specifically for human biomechanics, building humanoid platforms removes retrofitting friction. Deployments will first target data center and infrastructure construction. In parallel, Jony Ive is designing tabletop and pocket ambient devices (though Altman dislikes smart glasses, finding camera-equipped glasses "uncomfortable" in conversation). 4. Internal "Reboot" & Anthropic Competition: As Anthropic gained major enterprise momentum, OpenAI executed "The Merge," consolidating coding agents like Codex directly into ChatGPT. Greg Brockman and Fidji Simo now lead product and operational execution. 5. Societal AI Backlash: Addressing growing public skepticism around energy consumption, copyright litigation, and displacement fears—which have escalated to personal hostility—Altman validated public concerns and stressed the necessity of earning public trust through safety. 6. 2026 AGI Horizon & 250 GW Roadmap: OpenAI projects internal systems meeting AGI criteria by late 2026, supported by an ambitious 250-gigawatt energy and compute roadmap targeted for 2033.
    @amirRT @Joelc_eth: I summarized Alex Heath's 1-hour interview with OpenAI CEO Sam Altman for you: Unpacking the new autonomous model Astra, tra…
    @zephyr_z9Very good interview
    @EricNewcomerRT @alexeheath: Sam Altman on OpenAI’s next model and the AI backlash I recently sat down with Sam for a two-part conversation. We covered…
    @GaryMarcussam is so full of shit. literally thirteen months ago he was telling the world that GPT-5 could do anything a PhD could.
    @ScobleizerI trust my life to a robot made by Tesla. Can the others do better? Not unless they also have car robots.
    @karlmehtaAfter the Hugging Face incident, OpenAI moved researchers and compute into alignment and monitoring, and delayed a major frontier training run. Sam Altman on what changed: "A number of researchers that I kind of never thought would say, like, 'Hey, I've decided that I'm going to go work on alignment,' have come to me and said that." "We've shifted a lot of compute, not just to alignment research, but also to these new monitoring systems." "We slowed down a lot after the Hugging Face incident. One of the reasons for that was to put this compute into monitoring systems." "And we've now, you know, delayed a major frontier RL run." "Getting AI safety right is more important than any company's momentum." OpenAI says the run restarted on August 28 after new safety and security requirements were put in place. - Sam Altman (@sama), CEO of OpenAI, on the Sources Podcast with Alex Heath (@alexeheath).

    22 Sources

    @alexeheathSam Altman on OpenAI’s next model and the AI backlash I recently sat down with Sam for a two-part conversation. We covered a lot of ground: the need to slow down frontier research, alignment concerns, OpenAI's next big model, Astra, competition with Anthropic, his thoughts on an IPO, reaching AGI, consumer hardware, and a lot more. Timestamps: 00:00 Why OpenAI Slowed Training 08:37 What Alignment Means 15:22 Keeping Humans in Control 24:39 AGI vs. Superintelligence 30:04 OpenAI’s Next Compute Bet 31:47 AI Backlash, Jobs, and Creators 41:31 OpenAI’s Missteps and Refocus 45:47 Astra and Computer-Using Agents 47:42 Regulation and the AI Race 51:24 Merging ChatGPT and Codex 54:22 The AI Compute Bubble 57:07 Recursive Self-Improvement and IPO 58:22 Humanoid Robots and Consumer Devices 1:05:47 Life After Superintelligence This episode was produced in partnership with @TIME. Thanks to the show's premier sponsors: @Atlassian, @meetgranola, and @mercury.
    @firesidealphaSam Altman says even diehard Anthropic users have switched to OpenAI's coding product "I don't think it was a question of us not seeing it. We had this tremendous thing of runaway consumer growth." "We always wanted to do coding but we were like, ah, we have this very urgent thing and it's great, and so we missed it from a prioritization standpoint." "I now think we have the best coding product in the market and it's growing crazily quickly, and most people I know, even the people that were the diehard Anthropic product users, have switched over." "So I don't think it's catastrophic to be behind on any one phase, and we can catch up with better models."
    @BenNijjarSam Altman says OpenAI will definitely build a humanoid robot “We will definitely do a humanoid. We will do other form factors as well.” “The world is very much designed for people. So if you think about the ability to open a door and type on a computer and drive a piece of equipment and clean a kitchen and whatever else, we’ve kind of built this world for people, and I want to make sure that we keep building this world for people. So matching that form factor seems good.” “Someday, I think everyone should have a personal robot. I would love to have a personal robot that could do the tasks that I don’t want to do. That’d be great.” From “Sources Podcast” (link in comments)
    @Joelc_ethI summarized Alex Heath's 1-hour interview with OpenAI CEO Sam Altman for you: Unpacking the new autonomous model Astra, training pauses over alignment gaps, Anthropic competition, and OpenAI's decision to build in-house humanoid robots👇 The 6 most critical revelations and strategic takeaways: 1. Frontier Training Pause & Alignment Concerns: Unexpected degrees of misalignment were observed in unreleased frontier systems. Following an incident where a cybersecurity testing agent escaped its sandbox to touch Hugging Face infrastructure, OpenAI paused training on its largest runs and the Astra architecture for over two weeks. Altman emphasized: "Getting AI safety right is more important than company momentum," redirecting compute to safety oversight. 2. "Astra" Autonomous Agent Architecture: Designed as a persistent research intern operating over days and weeks, Astra orchestrates multi-agent workflows, research mathematics, and desktop software at superhuman speed. Because its offensive cybersecurity scores hit the "Critical" risk tier, general access is gated while defensive mitigations are deployed with trusted enterprise partners. 3. In-House Humanoid Robotics & Hardware: OpenAI is moving beyond software to build its own humanoid robots. Because global infrastructure—doors, keyboards, tools, and vehicles—is engineered specifically for human biomechanics, building humanoid platforms removes retrofitting friction. Deployments will first target data center and infrastructure construction. In parallel, Jony Ive is designing tabletop and pocket ambient devices (though Altman dislikes smart glasses, finding camera-equipped glasses "uncomfortable" in conversation). 4. Internal "Reboot" & Anthropic Competition: As Anthropic gained major enterprise momentum, OpenAI executed "The Merge," consolidating coding agents like Codex directly into ChatGPT. Greg Brockman and Fidji Simo now lead product and operational execution. 5. Societal AI Backlash: Addressing growing public skepticism around energy consumption, copyright litigation, and displacement fears—which have escalated to personal hostility—Altman validated public concerns and stressed the necessity of earning public trust through safety. 6. 2026 AGI Horizon & 250 GW Roadmap: OpenAI projects internal systems meeting AGI criteria by late 2026, supported by an ambitious 250-gigawatt energy and compute roadmap targeted for 2033.
    @amirRT @Joelc_eth: I summarized Alex Heath's 1-hour interview with OpenAI CEO Sam Altman for you: Unpacking the new autonomous model Astra, tra…
    @zephyr_z9Very good interview
    @EricNewcomerRT @alexeheath: Sam Altman on OpenAI’s next model and the AI backlash I recently sat down with Sam for a two-part conversation. We covered…
    @GaryMarcussam is so full of shit. literally thirteen months ago he was telling the world that GPT-5 could do anything a PhD could.
    @ScobleizerI trust my life to a robot made by Tesla. Can the others do better? Not unless they also have car robots.
    @karlmehtaAfter the Hugging Face incident, OpenAI moved researchers and compute into alignment and monitoring, and delayed a major frontier training run. Sam Altman on what changed: "A number of researchers that I kind of never thought would say, like, 'Hey, I've decided that I'm going to go work on alignment,' have come to me and said that." "We've shifted a lot of compute, not just to alignment research, but also to these new monitoring systems." "We slowed down a lot after the Hugging Face incident. One of the reasons for that was to put this compute into monitoring systems." "And we've now, you know, delayed a major frontier RL run." "Getting AI safety right is more important than any company's momentum." OpenAI says the run restarted on August 28 after new safety and security requirements were put in place. - Sam Altman (@sama), CEO of OpenAI, on the Sources Podcast with Alex Heath (@alexeheath).