
Getting the size of the market right doesn’t tell you who captures the most value. Introducing AI Perez: our attempt to make sense of shifting bottlenecks, and who benefits, through Carlota Perez’s framework. ai-perez-curve.vercel.app

tech investor & writer. 'truth will set you free.'

Getting the size of the market right doesn’t tell you who captures the most value. Introducing AI Perez: our attempt to make sense of shifting bottlenecks, and who benefits, through Carlota Perez’s framework. ai-perez-curve.vercel.app

One of the most requested machines are now live on @daytonaio 🍎
Your agents can finally use a Mac. macOS Sandboxes are now live on @daytonaio: M4, M4 Pro and M5 Pro. Spin one up at app.daytona.com

If you're a startup getting serious about security after the recent breaches, here's where I'd start: > Enforce SSO + MFA everywhere. > Get secrets out of code and rotate exposed credentials (@infisical can automate rotation for you). > Audit access and enforce least privilege, with just-in-time access for anything sensitive. > Replace long-lived credentials with short-lived ones (look up Infisical dynamic secrets and add OIDC where it makes sense). > Lock down CI/CD pipelines and pin your dependencies. > Stop giving AI agents unrestricted access to secrets (see Agent Vault). > Set up audit logs and configure a SIEM.


We made @theinformation list of top 50 promising early-stage startups for 2026 🎉 Thank you to everyone who runs, builds on, and talks about Ollama. theinformation.com/articles/…






For most of computing history, 3d was the most expensive kind of content. Someone had to build it by hand, one polygon at a time. Since 2022, AI has cut the time to generate a rough draft by about four orders of magnitude, from 90 minutes to under a second. A draft now costs well under a dollar. So if drafts are nearly free, where does the value go? ai-topology.vercel.app

just goes to show how many computers agents need
We just crossed $100M in run rate. Four quarters ago, we were at $1M. Then this year alone, we grew 37.5x. Not long before that, we did a hard pivot- handed our old customers to competitors, rebuilt from scratch, and relaunched with a few months of runway left. The bet was: models think on GPUs but work happens on CPUs. So every agent that writes code, opens a browser, or runs an RL job needs a composable computer that's isolated, stateful, and ready in milliseconds. $1M to $100M in about a year puts us in the same cohort as the likes of Anthropic and Cursor. It’s humbling to share that lane with companies we’ve learned so much from.



agreed
infisical is prob one of the least hyped (and underrated) startups out there rn not investment advice but if i had the moolah, i'd invest a few more underrated ones in my mind too
