Follow builders, not influencers.
TL;DR
Weil folded OpenAI for Science into core teams, while Google split Flow into music-making and Josh Woodward added remix control. Albert and Peter Yang showed Claude Design turning taste into production-grade assets, and Levie, Ryo Lu, and No Priors all argued AI wins when it serves workflows, not replaces them.
He says today’s his last day at OpenAI, where he helped launch OpenAI for Science and then moved from product into research. The big takeaway: he thinks accelerating science is one of the most positive things AGI can unlock, and now that effort is being decentralized across the company.
They’re expanding Flow beyond images and video with Flow Music, a standalone site for creating, sharing, and remixing original songs from natural-language prompts. It’s the same “type what you imagine, get a finished asset” pitch — now for tracks and playlists.
He says anyone with a vision can now produce high-quality designs with a little help from Claude. That’s the bigger story: Anthropic’s tools are pushing design from specialist craft toward something founders and product teams can iterate on fast.
He says enterprise platforms have to go headless because agents will use them 100X more than people do. That flips the old seat-based model: humans buy seats, agents drive consumption, and systems of record like Salesforce or Box can unlock way more use cases than they ever did before.
ProducerAI has been rebranded as Google Flow Music, with the same goal: give creators more control over their tracks. The new remix feature is the real hook here — more hands-on editing, less black-box generation.
The best agent isn’t just about a stronger model — it’s the combo of a solid harness, good models, and the ability to run anywhere. That’s a very Cursor-shaped take: the product wins by wrapping the model in the right workflow, not by betting on raw intelligence alone.
He says a feature he’s been using is especially good at anticipating small but important improvements to an app. That’s a very Replit-ish take: the best AI tools don’t just generate code, they nudge builders toward cleaner products and faster iteration.
He says Anthropic Labs is on a generational run, pointing to Claude Code, Skills, Claude Cowork, and Claude Design as the latest proof. The takeaway: Anthropic is no longer just shipping models — it’s building a full AI work suite around Claude.
He’s showing off Claude Design as a legit multi-format builder: videos, slides, websites, mobile apps, even a design system. The takeaway is pretty simple — AI is moving from “help me write” to “help me ship the whole thing,” which is exactly the kind of practical workflow Roblox product folks care about.
He says Fintool was already doing heavy reasoning before the models caught up, and calls it a “magical” product with a killer team now headed to Microsoft. He also gushes over a former Opendoor engineer starting a new company, basically saying he’d wire money before YC because the talent was that obvious.
He says every Claude model has its own level of helpfulness, and Anthropic’s Alex Albert explains why that matters for Opus 4.7. Every also published a full breakdown of how it performs on coding, writing, spreadsheets, and more — basically a practical read on where the model actually helps.
She says designing with code often works better than image models, and she’s been making most of her graphics in HTML instead of Nano Banana. The bigger point: output quality isn’t the same as good design, and LLMs aren’t “writers” just because they can spit out text.
AI wins when it serves workflows, not replaces them
The Takeaway: The real moat in enterprise AI is not model cleverness; it’s owning the workflow, context, and accountability.
Bill McDermott, CEO of ServiceNow and a veteran of Xerox and SAP, frames his career as a long lesson in customer obsession and disciplined execution. He started by buying a deli at 16, not because he was a natural entrepreneur, but because it consolidated multiple jobs into one and gave him control. That experience shaped his core belief: “the customer and the customer alone determines whether you win or lose.”
That same thinking drives his view of AI. He’s not buying the hype that language models will simply swallow enterprise software. Instead, he sees a split between answering questions and actually closing work. A model can suggest steps; a platform can route the case through HR, finance, legal, compliance, and risk until it’s resolved. His line is sharp: “People that run businesses understand that people make mistakes. They never will forgive software for making a mistake.”
McDermott’s bigger bet is that AI should amplify human ambition, not erase it. He wants ServiceNow to be the “AI control tower for business reinvention,” integrating hyperscalers, models, and enterprise systems while extending into security and operational technology. The philosophy underneath it is old-school but timely: opportunity matters, people matter, and the best technology is the one that helps humans do harder things faster without losing trust.
Anthropic launched Managed Agents to decouple agent infra, while Claude Code defaulted to xhigh effort and got a usage-focused upgrade. Rauch said agents need durability over clever prompts, and Swyx split AI engineering into slop vs rigor.
Rauch said teams were building their own design factories, while Steinberger called open-source AI security a full-time arms race. Masad priced OSS trust in compute, and Woodward shipped Gemini on Mac in 100 days.
Woodward said Gemini’s turning into a test-prep machine, Albert called Claude Code the whole workspace, and Cat Wu shipped a desktop control center with parallel sessions and review tools. Rauch also argued agent builders need elastic Postgres, not vibes.
Rauch said the moat moved from code to the code factory, while Levie argued every team now needed an agent wrangler. Cursor leaned into customizable multi-agent views, Replit added region controls, and No Priors backed Periodic Labs’ bet that AI could learn atoms by running experiments.
Amjad Masad said Apple’s 50th has turned into a PR disaster, while Aaron Levie argued agents would create more work, not cut jobs. Rauch pushed engineers into the customer hot seat, and Claude warned teams to harden security fast.
Thariq said Claude Code now handles TurboTax pain, while Rauch called microVM sandboxes the new compute layer. Aditya Agarwal pushed memory over loops, and Levie argued AI won’t shrink law—it’ll inflate it.
Claude pushed into Word with tracked edits, and Claude Code moved planning to the web with auto mode approvals. Garry Tan called agents the Altair BASIC era, while Aaron Levie warned software without a real API gets left behind.
Karpathy said free ChatGPT lagged while frontier coding models didn’t. Albert pushed cheap-to-smart escalation, Rauch said cloud infra went agent-native, and OpenAI’s next leap looked like autonomy—not chat.
Woodward gave Gemini a second brain with Notebooks, while Anthropic shipped Managed Agents to move Claude from prompt to production. Rauch called the web AI’s native OS, and Levie, Masad, and Shipper all bet agents will do the work, not the people.
Albert teased Anthropic’s Mythos Preview, Cat Wu juiced Claude Code’s CLI tricks, and Peter Steinberger patched CodexBar with 2 providers plus billing fixes. Levie said agents are eating knowledge work, while Nikunj Kothari preached retention over launch hype.
Levie said agents won’t erase work, just push it up a layer; Yang argued they’ll shrink teams, not ambition. Garry Tan flagged an unpatched file leak in Claude’s coding env, while Kothari called Anthropic’s revenue ramp absurdly fast.
Rauch said v0 now builds physics, not just UI, while Karpathy noted GitHub Gists have weirdly good comments. Levie argued AI efficiency creates more work, not less, and Tan called open source’s golden age.
Karpathy pushed “your data, your files, your AI.” Levie argued context beat raw model IQ in enterprise AI. Garry Tan said GStack kept shipping security fixes fast, while No Priors spotlighted Periodic Labs’ bet on atoms, not just text.
Claude plugged into Microsoft 365 everywhere, Swyx said Devin one-shot blog-to-code, and Peter Steinberger called out GitHub’s API as still not built for agents. Aaron Levie hit the context wall, while Garry Tan shipped a DX review tool from his own stack.
Claude landed computer use on Windows, Karpathy argued LLMs should build your wiki, and Amjad Masad pushed Replit deeper into enterprise sales. Peter Yang said Cursor 3 got out of the agent’s way, while Peter Steinberger warned AI slop was flooding kernel security with real bugs.
Steinberger called plan mode training wheels, while Thariq gave Claude Code a mouse-friendly renderer and Cat Wu showed sessions jumping phone-to-laptop. Masad framed Replit as an OS for agents, Rauch said Vercel signups compounded fast, and Anthropic’s infra tweaks swung coding scores by 6 points.
Levie said AI productivity hit the enterprise risk wall, while Weil argued proofs got cleaner, not just better. Agarwal floated public source code as the new prod debugging, and Data Driven NYC claimed one founder could run a company if agents handled the layers below.
Karpathy warned unpinned deps can turn one hack into mass pwnage, while Rauch and Levie said agents still need human guardrails and redesigned workflows. Meanwhile Claude Code got enterprise auto mode, Replit added built-in monetization, and Swyx spotted “Sign in with ChatGPT” already live.
Andrej Karpathy highlighted how LLMs can argue any side, suggesting we use it as a feature. Guillermo Rauch finally shipped his dream text layout, bringing his vision to life. Meanwhile, Amjad Masad claimed AI is democratizing app building and elevating top engineers.
Andrej Karpathy suggested leveraging LLMs' ability to argue any side as a feature. Guillermo Rauch turned text layout dreams into reality with Vercel's latest feature. Meanwhile, Amjad Masad claimed AI is democratizing app building, liberating top engineers for bigger challenges.