In this conversation, Robert Ta introduces GBrain, a self‑wiring knowledge graph built by Gary Tan (YC President & CEO) and released as a GitHub repo for OpenClaw/Hermes agent deployments. GBrain ingests 17,888 pages, 4,383 people, 723 companies, and 21 cron jobs in 12 days, extracting entity references and creating typed links without any LLM calls. It employs a hybrid search that boosts backlink‑based rankings and achieves strong benchmark results (P@5 = 49.1 %, R@5 = 97.9 %) on a 240‑page Opus‑generated prose corpus, outperforming a graph‑disabled variant by 31.4 points. The discussion then shifts to token‑cost considerations: monthly token budgets range from $40 (Haiku 4.5) to $1,000 (Opus 4.7). The architecture features a token‑matching front end for exploratory queries, “minions” (cached, database‑style agents) for low‑cost repeatable tasks, and queue jobs with deterministic routing rules that select the cheapest execution path based on usage history. This incremental automation mirrors Elon Musk’s staged automation approach, moving workflows from manual exploration to cached minions to fully automated closed‑loop systems.
Chapters
Gary Tan, who is the president and CEO of Y Combinator, built GBrain and released this earlier this year. GBrain is just a GitHub repository that bakes in a level of cross-session memory into OpenClaw or Hermes agent deployments. So some stats: 17,888 pages, 4,383 people, 723 companies, 21 cron jobs built in 12 days. That's GBrain. It is a self-wiring knowledge graph and every page, right, extracts entity references and creates typed links. So for Gary, it was stuff like attended this meeting, works at there, invested in X, founded this company, advises who. And it does all that with zero LLM calls. There's a level of hybrid search that backs things for backlink boosting of ratings, or I should say rankings, for search rankings. And he had an interesting benchmark. Benchmark side by side against the category, GBrain lands P@5, 49.1%, R@5, 97.9% on a 240-page opus-generated rich prose corpus, beating its own graph-disabled variant by 31.4 points. Thoughts on GBrain?
This is, so this is an extension of OpenClaw and Hermes usage. Um, I love that Gary Tan is using his GStack to then like implement GBrain, and that's all part of the, this, the sort of like offering here. So that's pretty cool that you have somebody, especially somebody with a design background, uh, building in this way and then releasing it. Now I think two things stuck out with me in like wanting to use the project myself. One was the token cost. There's a table that shows like how many tokens you should expect to use each month depending on what mode you're using it in. Yeah, there you go. Oh, here. Your token match budget goes up to $1,000 per month with Opus 4.7. Your conservative budget starts at $40 a month with Haiku 4.5. So, you know, we're using a lot of tokens here is the, you know, no doubt. And this gets into the architecture, which I thought was also really cool. If you scroll down, there's like a section on how it works and it shows this like really complicated sort of logic tree that's going to like try to, to like make sense of data using a lot of different techniques. And it's basically throwing compute at the problem, like a lot of compute at the problem. And so you have this kind of like big non-deterministic, uh, thing that's like trying to make sense of like the meeting that you just had or contact list or something like that. That part has been avowed with that, uh, precision and recall, uh, evaluations. And they actually created their own repo for those evaluations, which is also really cool. The thing is that that front part that uses like, you know, kind of a token matching approach to like categorize your information is just for exploring and kind of figuring out the workflows that, that might work for you. So you would be, you know, like typing queries into your agent. You'd be seeing like, is this useful? Do I wanna do this again? And if you do wanna do it again, and if it's a repeating thing, then you use this minions construct. And minions is kind of like, uh, a form of agent that is actually just like a database, uh, structure. Maybe it says store procedures that can run procedures over and over again for much less cost. So it uses much fewer tokens. So it's sort of like a cached, like this, this works and let's run it again type of job. And then you can further sort of graduate your workflows. You take them to minions and then you take them to like queue jobs. Oh yeah, they have a rule here, the routing rule. You were just looking at it. Deterministic means judgment and then subagents. This is a routing rule that actively figures out like, which of these, what's like the cheapest sort of system that I can run my request with, given my history with the system. So that's really cool. And that reminded me a bit of Elon's algorithm and the, you know, just gradual automation of things that get it to work first, see if it's valuable at all, and then move it incrementally into the automation phases until you have a fully automated system with closed loops. Cool project.
