Follow builders, not influencers.
TL;DR
Claude added interactive charts and Claude Code desktop with parallel sessions; Josh Woodward shipped Gemini conversation branching. Amjad Masad said static analysis lifted LLMs 90%+, while Aaron Levie and Guillermo Rauch framed agents and petabyte-scale hunts as the new battleground.
Interactive charts and diagrams are now in Claude Cowork, and it’s in beta for all paid plans. It’s a small but useful step toward making Claude feel less like a chat box and more like a real work surface for analysis and planning.
They argue that most bad products aren’t broken by one bad choice, but by a pile of reasonable additions nobody said no to. With AI dropping the cost of shipping to near zero, teams can now add features, concepts, and polish endlessly — and end up with a noisier, less coherent product.
He says the new ChatGPT agents are a huge shift because they can plug into tools and data directly, which is what knowledge work agents need to go mainstream. His Box example shows the play: enterprise content becomes a secure knowledge source, while agents generate answers and content on the fly across Box via MCP and CLI. The bigger takeaway is that headless platforms and enterprise agent builders just got a lot more interesting.
He says Vercel’s security team dug through nearly a petabyte of logs and found the attacker wasn’t just tied to the original Context.ai compromise — they were distributing malware to steal tokens and keys across providers. Vercel is now working with Microsoft, AWS, and Wiz, warning other victims to rotate creds while they keep shipping product fixes.
He says a papercut fix for conversation branching is rolling out to 20% now, with a wider ramp coming. Small feature, but it matters: branching makes it easier to explore alternate paths without nuking the original thread.
Replit says you can squeeze significantly better performance out of current-gen LLMs by pairing them with static analysis tools — in some cases, 90%+ better. He also pointed to Replit Agent now being callable from Gemini Enterprise, which is a neat distribution win for the product.
He says AI can generate the bulk of a thing, but the part that matters is the final 10% where taste and manual polish turn it into something worth keeping. That’s the real edge: everyone can ship faster, but only people who care about craft will make work they’re proud of.
He says the trick is to DRY up adjacent skills into bigger ones with branching params, instead of piling on lots of tiny resolvers. His take: shorter resolvers mean less context bloat, and that usually works better.
He says LLMs already write better code than most engineers, so the real human job is framing the problem and judging the output at the end. In his convo with Cora Computer’s Kieran Klaassen, he pushes the “AI sandwich” idea: models do the filling, while humans bookend the process and keep the repo learning via compound engineering. The punchline: one engineer can now ship like a small team if they know when to step in and when to let agents run.
GPT-Image-2-Thinking isn’t just a better image model — it’s an image agent with search, Photoshop, and self-review in the loop. The point: once generation takes minutes, the winner is the system that can iterate, composite, and correct itself, not the one-shot model. He says this is the text-to-image version of how Gemini Flash Vision broke image-to-text benchmarks with agentic looping.
Claude Code desktop adds parallel sessions and in-app review tools
Lead: Claude Code’s redesigned desktop app is built for parallel agent workflows, adding a new sidebar for managing multiple sessions, drag-and-drop workspace layout, an integrated terminal and file editor, and faster review tools.
Numbers:
So What: The update is aimed at developers orchestrating several tasks at once—refactors, bug fixes, and test-writing passes—without leaving the app. You can run sessions across repos, branch into side chats, inspect diffs, edit files, run tests, and preview HTML or PDFs in one place. The company says the app is designed for “how agentic coding actually feels now: many things in flight, and you in the orchestrator seat.” For teams, plugin parity with the CLI and centralized management means the desktop experience should slot into existing workflows with minimal friction. If you’re already using Claude Code, update the app and start consolidating session management, review, and shipping into a single workspace.
Humans should frame and polish; AI should do the middle
The Takeaway: The winning model isn’t full automation — it’s putting humans at the edges where judgment, taste, and ownership matter most.
Key Insights
The Story
Kieran, GM of Quora and creator of the Compound Engineering framework used inside Every, built his approach by asking a practical question: how do you get AI to do better work faster without flattening the human out of it? His answer is a workflow with planning, execution, review, and a “compound” step that stores lessons back into the repo so agents improve over time. But the deeper insight is where humans belong. Early on, humans should be deeply in the loop for ideation and problem framing; later, they should step back and let the model execute. At the end, they should come back with taste. As he puts it, “the beginning and the end, and the middles can be automated pretty well.”
That’s why he pushes back on the idea that humans should always stay involved. The real leverage is knowing when to think hard and when to hand off. He compares it to music: the middle is practice, but the performance is where something alive happens. The same is true in software, design, and writing. If you want the output to be yours, it can’t be fully outsourced. The bar keeps rising, so the human edge becomes less about grinding through tasks and more about making something that “feels great.”
Altman said OpenAI wanted you swimming in AI—and GPUs. Masad pushed for a fairer software market, Levie said enterprise agents needed humans to actually land, and Shipper showed agents could now read voice notes.
Rauch said delete isn’t rotation, Levie argued agents need operators, not just users, and Steinberger kept OpenClaw pushing AI into real workflows. Shipper backed two-agent setups, while Claude warned teams to harden security now.
Rauch said an AI-accelerated attack exposed Vercel’s weak link, while Kothari warned AI will supercharge attacks too. Garry Tan called Claude Code the new app factory, and Peter Yang noted agents still flaked on boring cron jobs.
Rauch said design was becoming autonomous, not just a tool. Steinberger made CodexBar safer, faster, and lighter; Anthropic added Auto Mode to Claude Code and showed benchmark scores can swing with eval infra. Levie warned AI agents would force constant rewrites.
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.
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.