Deepnote

Going from 1 employee → 2 employees right now makes you 10x less productive. Building alone is now absurdly productive. Adding just one team member breaks everything. Everything runs from your laptop using your personal credentials. Working in a team suddenly means figuring out permissions, safe edits, coordination, and where to run an always-on agent. As a thought experiment, I built a viable startup alone in a few weeks with Claude Code. Adding a second person made everything 10x slower. My edits broke their app. The agent used my email, API keys, and database credentials. I couldn't hand any of it over. There was no simple solution. The “company in a box” products I tried had the same problem. Whoever solves that could replace Notion, Linear, CRM, and pretty much every other tool in “productivity” category.

Permanent interfaces for temporary work are so 2025. To cut next quarter’s software spend, I need a chart, an editable budget, and an approval queue. When the job is done, the interface can go. The decision logic and context can be codified. This is what I see coming in the near future, and both OpenAI and Anthropic have already announced a version of this. Software and UI will assemble around tasks. Every SaaS CEO should prepare for that.

Reverse Engineer Anything is actually insane. I ran it on Deepnote over the weekend. The results blew my mind. I open-sourced it all here: github.com/deepnote/deepnote We're so back, it's so over.

I am trying out a new way of working with agents. I give them business context, a clear goal, and authority over two-way door decisions. If we can undo it, they shouldn’t wait for me; they should keep going. If the agent thinks it needs to make a decision and it’s a one-way door (e.g., changing pricing, buying tools, signing contracts), it goes into a queue I review each morning. They keep working on something else until I review it. If the agent thinks the decision is a two-way door (e.g., choosing a new landing page layout, rewriting internal policies), it makes a judgment call. The decision gets logged, and I review it next morning. The agent always works, never gets blocked. The number of things I’m building this way is crazy. This is the way to build products and companies in 2026 and beyond.

Six analysts serve 1,000+ people at a company whose data stack grows with each acquisition. They unify the data and ship actionable insights with AI in Deepnote. Atolls grows by buying other companies. In practice, that meant stitching CSVs together in Jupyter and sharing screenshots in Confluence. So, they moved those workflows into Deepnote. What changed: 1. Data sources sit side by side in one workspace. 2. Analyses reach stakeholders the same day, down from 1–3 days. 3. Self-serve data apps save them 8–10 hours a week on recurring requests. Now the team has more time for deeper analysis and has already shipped their first machine learning models into production. Huge thanks to the Atolls team for trusting us with their critical workflows.

2026 really is the year of agents. Mostly federal.

Tibo: “Here’s another limit reset.” Me: * asks Astra for one code review * Codex:

We’ve moved on from vibe coding to agentic engineering. Vibe coding = you don’t write the code, you don’t read the code. Code has security holes and hallucinations. Can be used for prototypes, nothing else. Agentic engineering = you don’t write the code, you don’t read the code. But the code is better than code written by a human. You can guarantee it’s correct. You can build large systems.

Every week, there is a new startup building its own harness. But the best harnesses will be built by the labs. They control the models and can optimize both together. They also have a strong incentive to build the best harness. So if you’re building a harness, we’re now back to startup 101: pick a niche. Not too small so there is enough revenue. Not too big so you don’t get outcompeted by a giant with more resources and better distribution.

@OpenAI just launched the Data agent in ChatGPT Work, with @DeepnoteHQ available as a native connection. Connect Deepnote to bring the data sources and context your team already uses into the conversation, then explore your company’s data, build interactive data apps, or take action. To get started, add the Data Plugin and Deepnote in ChatGPT Work, and start your exploration: vist.ly/5iin8 Special thanks to Charles Wang, Steve Imm, Oliver C., and Shaina Li for the collaboration.

In a year, you won't know which model you're running. Fifteen years ago, people looked for a Pentium or Celeron sticker before buying a laptop. Then chips got fast enough that the sticker didn't matter anymore. Models are heading the same way. Most tasks don't need the frontier, and routing to a cheaper model saves a lot of money. As a result, I believe that by the end of the year we will no longer pay per token but per some obscure intelligence unit. ChatGPT is currently testing hiding models from you, with Sol as a default and lower-capability models being routed to. I think this will soon become the new norm for all providers. Do we need one more black box in AI?

OK guys, be honest, who's distilling models en masse again?

It’s getting harder to tell a larger model from a smaller one. Once the output is comparable, speed becomes the new differentiator. The intelligence gap is closing, and at some point, a smarter model isn’t worth waiting for. Everyone says that Google is too far behind to catch up, but I think they are building for this.

I was on a plane with bad wifi and 15% battery left, so I told my agent to hurry up. Best experience I've had with it in months. It makes sense. Benchmarks are optimizing for the highest score, not my time. So every model is incentivized to lint, typecheck, spin up a browser, take screenshots, typecheck again. Great for fixing bugs and maxxing out benchmarks, not so great for building. Video credit: @blended_jpeg

POV: my four agents merging into main at the same time

Should agents decide their own workflow? Besides loops, maybe it’s one of the biggest debates in AI engineering right now, and we're having it inside Deepnote too. Non-deterministic won’t fly in enterprise, which is where we see a lot of adoption: some workflows need guaranteed, auditable execution patterns. Good thing is, we’ve architected Deepnote to enable both types of workflows. Notebooks are actually pretty good at that → isolated environments where non-deterministic runs can happen (and plain English instructions can be beautifully formatted) + deterministic behavior via code, with a clear sequential/reactive pattern.

Should we stop building UI entirely? A legit question I catch myself asking as I build Deepnote. We attracted strong demand initially due to having a much better UI vs incumbents. I'm increasingly unsure if any software company should keep shipping more screens. Work already starts in Claude Code, Codex, or Cursor. For better or for worse, the chat is the front door now. But whatever gets built there still needs a place to connect to data, deploy, run, collaborate, and govern access. Giants like Salesforce are embracing being headless for agents. If you think about it, each SaaS product eventually boils down to being an API or a database. Maybe we should take this into account as we build out the companies.

AI vendor lock-in is dying before it ever shipped. Every company I talk to is building the same thing: a CRM, an internal data analyst, a productivity tracker. Think of this as an operating system custom-made for your business — on top of whatever model is cheapest that week. Whoever hosts this operating system will have the moat. Right now, that's increasingly the companies themselves. For OpenAI and Anthropic, this is both a threat and an opportunity. If companies build this first, they can switch to another model anytime, including the open source models. If OpenAI/Anthropic build this first, they’ll finally get some vendor lock-in.

Agent Harness > Model The model leaderboard reshuffles every week. Teams that hardwire themselves to one lab spend every release cycle re-evaluating their whole stack (and life choices). Teams that own their harness can just flip the switch. We felt this at Deepnote. We once moved back a model generation because A/B tests with real users preferred the old one. The swap took minutes. Everything that made the agent useful stayed: the context layer, permissions, traces. That's the harness. And it's where the compounding happens. It knows your schemas, your permissions, your definition of done. If configured correctly, it can even outperform the vanilla harnesses by the frontier labs on ARC-AGI-3 (as shown by the @PrimeIntellect team). The labs know this. It's why every one of them now ships compound systems, and not just API-driven models.

At @DeepnoteHQ, we're very data-driven. Also, the data in question: