Forward Future Live — Matt Schumer, OpenClaw Explosion, Warp Oz & Shield AI

Matthew Berman and Nick Wentz. Forward Future Live. February 13, 2026. Detailed Summary.

This is a 1 hour 43 minute live show hosted by Matt Berman and Nick Wentz, featuring three guests: Jaden Clark from the Motts podcast covering San Francisco AI culture, Zach Lloyd the founder and CEO of Warp, and Armor Harris the SVP and GM of Shield AI. The episode covers Matt Schumer's viral article, the explosion of OpenClaw, Spotify's zero-code development, simulated societies, Warp's new AI agent platform Oz, and autonomous fighter jets.

Section 1. Matt Schumer's 80 Million View Article.

The show opens discussing Matt Schumer's article on X called "Something Big is Happening" which hit 80 million views. Matt and Nick break down why it went viral. Matt says it was a mixture of hype, fear, and realism. It spelled out what people inside the AI bubble were feeling but couldn't communicate to outsiders. It was specifically written for insiders to send to their friends, family, and colleagues.

Both hosts sent the article to their wives. Matt says his wife read it right before bed and said "Well, now I'm not going to be able to sleep." The article had a deep effect on people because it encapsulated the step change that happened in the last few weeks, specifically with Opus 4.6 and GPT 5.3 Codex. The confidence level in sending AI agents off to complete tasks autonomously has fundamentally changed.

Nick identifies three ingredients that have exploded his AI usage: the new frontier models, particularly Opus 4.6, the new code generation capabilities of GPT 5.3, and OpenClaw as the third ingredient that completely unlocked applying AI across different parts of their lives and business.

Section 2. Automating Their Own Business.

Matt and Nick share that they are actively automating large parts of their media business. Matt and his team member Jonah had been doing video research and organising thoughts together, and they've now automated that entire workflow. But the counterintuitive finding is that they're not doing less work. They're busier than ever, thinking about so many other things they can now do.

Matt says this is an indicator for what's to come. If you're fearful about being automated away, as long as you have adaptability, you're going to be fine. There's so much to do out there. Nick reinforces this, noting that Microsoft's deputy CTO Sam Salaice said the same thing: everyone's calling for a white-collar bloodbath, but he's never been busier and has never had more projects going on.

They identify a new problem: attention saturation. Matt says he's struggling to keep up with all the things his agents are doing for them. Orchestrating all these agents and managing the outputs is itself becoming overwhelming.

Section 3. OpenClaw's Cultural Impact.

The hosts discuss OpenClaw's broader impact on the AI ecosystem. They note that OpenClaw for Slack hit 1 million ARR within three hours of release. About a hundred businesses have been built in the last two days on top of OpenClaw services, including hands-on setup services.

Matt makes an insightful comparison to the early internet. When the internet first arrived, people just took newspapers and magazines and put them online. They applied the old form factor through the lens of the new technology. It wasn't until people started experimenting with what else was possible that things got interesting. He says OpenClaw did the same thing for AI. Before, most people were using AI as a Google search alternative. OpenClaw opened people's eyes to what's actually possible: deeply integrated personal assistants living on your machine that can accomplish virtually any task done in front of a screen.

The key innovations: the level of integrations possible, the fact that it lives locally on your machine giving a sense of control, and that it shows up natively in chat apps like WhatsApp and Telegram where you just tell it what to do and it figures it out.

They also discuss the news that OpenClaw's creator may be joining a major frontier lab, possibly after conversations with Zuckerberg and Altman. Matt views this with mixed feelings. He sees the creator as a champion of open source and frontier experimentation, an underdog figure whose independence would be lost at a big company. The creator has said the project will stay open source, but the hosts note that big companies aren't historically great at maintaining open-source commitments when shareholder interests conflict.

Section 4. Spotify's Zero-Code Development.

Spotify announced that its developers haven't written a single line of code since December 2025. They're using an internal AI system called Honk, powered by Claude, and have shipped 50 plus features. Matt notes this is becoming more and more common, whether it's the Claude Code team or developers saying they wrote entire products with AI assistance. Developers are the first to realise they're being automated, and they're responding by getting more work done than ever, shifting from writing code to orchestrating AI.

Section 5. Simile and Simulated Societies.

The hosts discuss June's new company Simile, which raised a significant round to build AI simulations of human behaviour. June was the author of the famous Stanford "human simulacra" paper from two years ago, where 100 AI agents placed in a virtual town developed their own personalities, relationships, and emergent human-like behaviours. Agents would build friendships, invite each other to birthday parties, and introduce their friends to each other.

Simile's pitch: pilots don't train with real passengers, surgeons don't practice on real people, actors don't rehearse with real audiences. Yet our most consequential societal decisions in government and policy are pushed straight into production. Simile aims to build AI simulations of society to test policies before they're released to humans. They're already partnered with CVS to simulate customer purchasing decisions. The really interesting application is at the societal level, testing what happens when you change a law or policy and forecasting the impact.

Section 6. Jaden Clark on San Francisco AI Culture.

Jaden Clark, an Australian living in San Francisco, hosts the Motts podcast covering niche SF technology culture. He describes himself as "anti-hype," always stepping back to assess whether the buzz is justified or just crazy posting. His answer: it's almost always a bit of both.

On the white-collar bloodbath narrative: Jaden pushes back strongly against Dario Amodei's predictions. He argues that until AI can one-shot tasks with 100 percent accuracy, and until you can hold an AI liable in court, humans must remain in the loop. This is especially true for accounting, legal work, and anything where accuracy is critical. He favours augmentation over automation and thinks jobs will be augmented for years to come.

He raises an important messaging point: when someone like Dario predicts 20 percent unemployment in 1 to 5 years, Jaden's reaction is "my brother in Christ, you're making the sandwich." He thinks the messaging from AI labs matters enormously and that poor messaging could lead to blunt regulation, possibly as early as the midterms. Data centres are already being targeted by both political parties over electricity and water usage.

On San Francisco specifically: the energy is stronger than ever. It feels like another gold rush. The unique thing about this particular AI boom is that the major labs have headquartered in the city itself, not the peninsula like the FANG companies. There's more venture capital, more funded projects, and talented founders being funded. He also got to pilot a humanoid robot in a cage fight, which he describes as one of the craziest things he's seen recently.

On China versus US sentiment: the contrast is stark. China focuses on the utilitarian aspect of AI, while the US focuses on godlike intelligence that will take over everything. China's framing has had a much more positive effect on public reception.

Section 7. Zach Lloyd and Warp's Oz Platform.

Zach Lloyd, founder and CEO of Warp, introduces Oz, their cloud agent orchestration platform. Warp started as a terminal replacement, the first terminal application built with a GPU-rendered canvas rather than a text grid. It's grown to hundreds of thousands of active developers and is now profitable.

The big insight: as AI transforms software development, the terminal is actually becoming more important, not less. Developers are spending more time in terminals because AI agents write code through command-line interfaces. More code being written means more code being deployed, which means more terminal usage.

Oz is designed for a world where agents are running in the cloud doing work autonomously. It provides an interface to manage, orchestrate, and monitor those agents. Zach describes it as a "mission control for agents." The key challenge they're addressing: how do you maintain oversight when you have many agents working simultaneously? The attention saturation problem that Matt and Nick described is exactly what Oz aims to solve.

Zach was cautious about the agent hype cycle, noting that many people are doing impressive demos but the gap between demo and reliable production use is still significant. He thinks the terminal will become the primary interface through which humans interact with AI agents, not chat windows.

Section 8. Armor Harris and Shield AI's Autonomous Fighter Jet.

Armor Harris from Shield AI discusses the XBAT, an autonomous vertical takeoff and landing fighter jet. Shield AI is building autonomous military aircraft designed to maintain deterrence through what he calls asymmetric means.

The key argument: in modern warfare, you don't want to be sending manned aircraft into the most dangerous situations. Autonomous systems can take on the highest-risk missions while keeping human pilots safe. The XBAT is designed to be significantly cheaper than traditional fighter jets while being expendable in high-risk scenarios.

Harris is careful to emphasise the need for human oversight in autonomous weapons systems. His position: you need humans in the loop for any lethal decision-making. The goal isn't to remove humans from warfare entirely but to give militaries tools that can operate in environments too dangerous for human pilots while maintaining human command authority.

He discusses the broader trend of AI in defence, noting that the Department of Defence is increasingly looking at autonomous systems as a way to maintain military superiority without the enormous cost of traditional manned platforms. A traditional fighter jet costs over $100 million. An autonomous system could cost a fraction of that while performing many of the same missions.

The philosophical tension: how do you prevent the dehumanisation of warfare when machines are doing the fighting? Harris argues that paradoxically, autonomous systems could actually make warfare more humane if they reduce civilian casualties through more precise targeting and if they keep human pilots out of harm's way. But he acknowledges this is an active debate with no easy answers.

Section 9. Closing Themes.

The episode wraps with reflections on the pace of change. Matt reiterates that the last few weeks have felt different from anything before. The combination of frontier models reaching new capability thresholds, OpenClaw demonstrating what's possible with deep integration, and the sheer speed of businesses being built on top of these tools points to an acceleration that's hard to fully comprehend.

The recurring theme across all guests: we're at an inflection point, but it's not the apocalyptic narrative. Developers are busier, builders are more energised, San Francisco is more alive than ever, and the real challenge isn't losing jobs but managing the overwhelming number of new possibilities. The advice: stay adaptable, get your hands on the frontier models, and build something.