Why Recursive Self-Improvement Hits a Wall | Ramez Naam

Melvin Johnson, VP of research leading Gemini post-training at Google DeepMind, joins MTS to unpack the rapid rollout of the Gemini 4 model family—Argon, Carbon, and today’s workhorse release, Hydrogen. He breaks down how Hydrogen outperforms flagship models on key benchmarks, why Google’s vertical integration with TPUs and massive distribution keeps frontier labs ahead of open-weight models, and where agentic consumer products like Search AI Mode are heading. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS chronicles the singularity, live. Monitoring the situation of tech and AI.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (00:46) The Gemini 4 Model Family (02:52) Google’s Structural Advantages And Strategy (04:55) Frontier Labs Versus Open-Weight Models (07:34) Future Consumer Agentic Products–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Vercel AI Product Lead Harpreet Arora shares enterprise adoption trends from Vercel’s AI Gateway, revealing how teams are shifting toward Anthropic models, decision models, and voice workflows. He also discusses the slowdown in open-weight model quality improvements relative to frontier labs, NVIDIA’s role in the open-source ecosystem, and predicts video models will become reliable for repeatable workflows within a year. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.--MTS chronicles the singularity, live. Monitoring the situation of tech and AI.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (01:05) Enterprise AI Trends on Vercel Gateway (02:44) Anthropic Beats OpenAI With New Teams (03:32) NVIDIA and Open-Weight Model Setbacks (09:00) Do Enterprises Care About US Runtime? (11:08) Anthropic vs. OpenAI Enterprise Use Cases (13:04) Voice Models Hit Rapid Team Adoption (13:54) Video Models and Future AI Predictions–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Horizon3 co-founder and CEO Snehal Antani discusses the launch of the company’s Adversarial Intelligence Frontier Lab and how AI agents are changing offensive and defensive cybersecurity. He explains why frontier AI models fail against basic cyber deception, why private behind-the-firewall training data is the real moat over compute, and why attackers rely on machine-speed execution of basic flaws rather than zero-days. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.https://x.com/snehalantani--MTS chronicles the singularity, live. Monitoring the situation of tech and AI.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (01:07) Adversarial Intelligence Frontier Lab Launch (02:12) Why AI Cyberattacks Are Accelerating (04:13) Compute vs. Cyber Training Data (06:51) Why Frontier Labs Lack Data (10:14) Horizon3 Hits $2B Valuation (11:44) Black October and Hacker Tactics (15:38) What’s Next for Horizon3–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Google DeepMind research engineer Ege Erdil explains why he is bearish on $1T+ AI lab valuations, comparing frontier labs to low-margin, high-CapEx airlines rather than traditional software monopolies. He also challenges short-timeline forecasts like the AI 2027 intelligence explosion thesis, arguing that AI progress remains bottlenecked by compute and data rather than self-accelerating software research, and estimates full automation of white-collar work is still roughly 20 years away. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS is brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (00:59) Do Trillion-Dollar AI Lab Valuations Make Sense? (06:15) Is AI a Winner-Takes-All Market? (09:43) Could Frontier AI Labs Go Bankrupt? (12:54) Best AI Trades: Memory Stocks and Nvidia (15:53) Why the 2027 Intelligence Explosion Won’t Happen (25:39) Evaluating the METR Report on AI Acceleration (28:00) White-Collar Automation Timelines and Questions for Buck–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Runway co-founder and CEO Anastasis Germanidis explains why the video generation lab open-sourced Praxis-1, its new world action model topping NVIDIA’s RoboLab 120 benchmark. He breaks down the race with China on physical AI, why predicting pixels trains real-world robot control, and his forecast for robot deployments moving into production by next year. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.https://x.com/agermanidis--MTS is a live news and interview show covering technology, business, politics, and culture as it happens.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠Interested in sponsoring the show? ⁠⁠⁠⁠⁠⁠[email protected]⁠⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (00:51) What Praxis-1 Means For Robotics (01:43) Why Video Pre-Training Controls Robots (02:57) Real-World Deployments And Office Robots (05:23) Why Open-Weight Robot Models Matter (07:43) The Race For US Leadership (09:34) Topping NVIDIA’s RoboLab 120 Leaderboard (10:57) When Generalist Robots Hit Production–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Former White House AI advisor Ben Buchanan details the geopolitical battle for AI dominance, how chip export controls are starving China of frontier compute, and what happens if the US faces superhuman AI by 2028. He explains the Biden administration’s behind-the-scenes strategy to keep the frontier lead and why computing power remains the ultimate bottleneck. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS is a live news and interview show covering technology, business, politics, and culture as it happens.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠Interested in sponsoring the show? ⁠⁠⁠⁠⁠⁠[email protected]⁠⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (01:44) Getting AGI Pilled Early (05:09) Inside The Export Control Debate (10:29) The Manhattan Project Scenario (13:05) Grading Trump’s AI Policy (18:34) Regulating AI Without Economic Damage (20:28) Preparing For Accelerated Timelines (22:05) Lessons From White House Tenure–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

ElevenLabs Field CTO Alex Holt breaks down their major Eleven Eleven Summit launches, including the new Interaction Engine, agent orchestration, and the NVIDIA Nemotron partnership. He explains why voice delivery is the hardest bottleneck in enterprise AI and forecasts that voice agents will replace traditional IVR by the end of next year. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS is a live news and interview show covering technology, business, politics, and culture as it happens.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠Interested in sponsoring the show? ⁠⁠⁠⁠⁠⁠[email protected]⁠⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (01:09) ElevenLabs Summit Major Product Announcements (02:23) Orchestrating Multi-Agent Enterprise Workflows (04:57) Closing the Enterprise Deployment Gap (07:08) Why Voice Agents Are So Hard (10:25) Building the Technical Interaction Engine (11:25) When Voice Agents Become the Norm (15:05) Transforming Healthcare With Scribe V3–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

ElevenLabs Chief Revenue Officer Ashley Kramer joins MTS at the ElevenLabs Summit to break down how voice agents are moving from enterprise pilots into full production. She shares real-world deployments across healthcare, telecommunications, and hospitality, including Clover Health’s Medicare rollout, KPN’s Dutch verification system, and Deutsche Telekom’s real-time network translation layer. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS is a live news and interview show covering technology, business, politics, and culture as it happens.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠Interested in sponsoring the show? ⁠⁠⁠⁠⁠⁠[email protected]⁠⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (00:46) Transforming Medicare With Clover Health (02:23) Forward Deployed Engineers For Enterprises (04:56) Why Users Will Prefer Voice AI (07:12) Global Expansion And McKinsey Alliance (09:27) Entering The Next Interaction Phase (10:54) Biggest Blockers To Enterprise Adoption (14:22) Moving Beyond Pilots Within A Year–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Planetary VC investor Ramez Naam pushes back on runaway recursive self-improvement models and short AGI timelines, arguing that empirical lab data shows steep diminishing returns rather than exponential feedback loops. He compares theoretical benchmarks to real-world task horizons at OpenAI and Anthropic, explaining why fully autonomous AI researchers remain constrained by data, taste, and verification bottlenecks. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS is a live news and interview show covering technology, business, politics, and culture as it happens.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠Interested in sponsoring the show? ⁠⁠⁠⁠⁠⁠[email protected]⁠⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (00:56) AI Futures Project 2027 Forecast (04:50) Dollars In Versus Dollars Out (10:31) AI Futures Cruxes And Benchmarks (12:17) Evidence Needed To Update (15:09) Why Forecasters Have Short Timelines (21:47) Updating On Recent Uplift Data (22:39) Superhuman Researchers Versus Generalized ASI–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.

Alberto Arrighi and Max Poff from Stack AI break down what it actually takes to push enterprise AI agents into production at scale. They discuss moving beyond simple chat demos, securing air-gapped deployments for defense contractors, and their timeline for embodied intelligence. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love.MTS is a live news and interview show covering technology, business, politics, and culture as it happens.Follow MTS: X | ⁠https://twitter.com/MTSlive⁠YouTube | ⁠https://www.youtube.com/@mtsituation⁠Spotify | ⁠https://open.spotify.com/show/4HUNNmV1pjV7CW6Lphm1rM⁠ Apple | ⁠https://podcasts.apple.com/us/podcast/mts/id1891088763⁠ Substack | ⁠https://mtslive.substack.com/⁠Interested in sponsoring the show? ⁠⁠⁠⁠⁠⁠[email protected]⁠⁠–Brought to you by:Mongo DB - Anyone can prototype an agent. Learn how to get them into production on September 30th in New York - register at https://www.mongodb.com/events/mongodb-local/nycChargebee - Billing and monetization infrastructure for the AI era. https://www.chargebee.com/Lovable - Turn ideas into software people love with Lovable. Start today at https://lovable.dev/ElevenLabs -  AI that communicates at human level across every channel and modality. https://elevenlabs.io/mtsFactory - A 24/7 system turning signals into production software. Visit https://factory.com/Neon (by Databricks) - Scale your startup on Neon. The Postgres backend for apps and agents. Built by Databricks. ⁠https://neon.com/mts⁠Today’s stream is sponsored by VCX, the public ticker for private tech. Head over to http://getvcx.com  to learn moreKong - The AI connectivity platform. Connect APIs, LLMs, agents, and systems with serious security and governance. https://konghq.comLarridin - Finally prove what your AI investment is worth. Head to https://larridin.com/ today to start measuring your AI’s bottom line.Blitzy - Autonomous software development for enterprise codebases. Ship 5x faster. https://blitzy.comArena - Measuring AI performance in the real world. ⁠https://arena.ai/⁠AdQuick - Make your brand a billboard. Out-of-home advertising as easy to scale as digital. https://adquick.comClickhouse - The real-time database that keeps up with all of it. https://clickhouse.com/–Timestamps:(00:00) Introduction (01:01) Roles in Enterprise AI Strategy (02:40) Enterprise Behavior Around Adopting Agents (05:15) Hardest Parts of Production Scale (07:08) Engineering Stack and RSI Realities (10:01) On-Prem Deployments and Security Trends (13:24) Unlocking Scale With Asana Integration (16:05) Embodied AI and 2027 Predictions–Note: This podcast is not investment, legal, or tax advice, and is intended for informational and entertainment purposes only. Hosts and guests may hold positions in the companies and securities discussed; do your own research before acting on anything you hear.