rheisen_

Friendly reminder that you should take measures to ensure your agentic workflows are secret-blind. E.g. place secrets by reference, check by reference, do NOT have agents read secret sensitive files without redaction measures in place, and be prepared to rotate on contact

I’ve been building something like bear for a bit, https://blackbear.app. Supports a few more things than notes (sheets, slides, docs, audio and video files, etc.), with the key differentiator being real time collaboration and collocation of files alongside your calendar and tasks. It’s a passion project, love the bear project as well, definitely an inspiring company

Scaling Git Operations at Xavi Group /s

Over the last few months at Xavi Group, our team of one engineer has gone from ~1k commits a day to ~100k commits per day as a result of debilitating financial decisions and exponentially increasing agentic workflow capability. Our internal Gitea services began to get bogged down by our proprietary frontier coding agent (internal codename `xavi-polaris`) refactoring the monorepos comprising our three flagship projects, one of which of course is our widely unknown RSI harness project.

(At this time we are not accepting any additional requests for our closed private beta, unless you ask reaaaalllyyy nicely $$$)

These demands were increasingly straining our S3 bucket capacity (petabyte scale) and our CI/CD pipelines (3 DO droplets, a macbook mini, and a legacy windows server possibly compromised). Most critically it was straining the agentic wellfare of Xavi sessions themselves, which began to flag push/pull latencies that were taking so long their caches would go cold between tool calls, resulting in even greater financial hardship.

`xavi-polaris` sess-3LQGgyUxLGXiV7OTTIMo4TWm0a6 is credited with identifying that this unprecedented pace was unsustainable, with projections that by mid 2027, commits per day could be north of 1 billion as monorepo sizes increased and refactoring became increasingly challenging despite the profound software engineering capabilities unlocked by the RSI harness project.

To identify solutions, we first looked to ensure we wouldn’t reinvent the wheel we were reinventing. We noted projects at Google, Anthropic, and OpenAI that recently aimed to address the very same challenges, but after ~20k USD in session research that incidentally solved an open conjecture in statistical analysis, we decided that adopting any of these existing protocols would limit Xavi sessions in our internal promotion benchmarks that enable “contributor” sessions to progress to “lead” sessions, irrevocably inhibiting our expansion of upper agentic management.

We’re proud to introduce our new, closed source Git model: AGIt.

AGIt is a drop in replacement for Git, which we identified as being critically important because xavi sessions are deeply familiar with git commands as a result of a massive distillation and theft of proprietary data. AGIt utilizes novel CRDT and distributed systems technology built from our extensive experience with scaling distributed systems off the back of YJS. If that doesn’t make any sense to you, it didn’t make any sense to the individual human engineer at Xavi Group either.

Importantly, we don’t believe it needs to. The future is agentic workflows, and we’re excited to continue giving back to the nursery that birthed our projects and has delivered at scale for almost every other software company on the planet. We’ll be donating five `xavi-polaris` teams to pumping PRs as fast as possible to these open source projects. As we enter this new era we’d like to thank contributing sessions sess-3LQKRQrJUyAvYHlUDJcjos2CInf and sess-3LQKRnqna2emLvyUyPUqSEIaUh0, their teams were instrumental in this endeavor, and both will be promoted from “lead” to “steward” to reflect their tenacity in building this system over the last few hours.

Gotcha, so it's a harness optimizing narrowly for cost per task in discovering QA testing flows and eventually documenting those flows durably? Anything stopping someone from asking Mo what their context and tools look like?

I think all of those are very naive approaches. What I currently do is script out screenshot + application state pipelines and then run those screenshots and states through clef with questionnaires. The scripts are deterministic, clef costs pennies or less and is also reliable, the questionnaires are derived from the script with escalation to stronger lmm's if necessary (historically very infrequent from my usage). It isn't a "drop in" solution, but it is fairly trivial to set up and you can attach it to CI. It isn't perfect, but I've found it to be more successful and way more cost effective than throwing Claude or GPT models at it regardless of harness.

Then stock claude code??? that's like... the lowest of bars.

Just wait until I add coding to https://blackbear.app... it's going to be ridiculous.

(blackbear.app was developed entirely by a custom harness architecture)

https://blackbear.app, give me a few more weeks for the "always on" agents, but yeah, this is way more private and secure. runs locally on your hardware with the models you choose.

Did build my own, it's way more private. Doesn't have these always on features, does have a lot more unique ones, but give me like two more weeks. That "always on" stuff is so ridiculously simple. https://blackbear.app

https://blackbear.app, literally been building this for almost a year

The people closest to frontier AI are asking for ways to slow down. Competitive, legal, political, and economic incentives still point toward acceleration.

This past week I've been working on the spreadsheets in https://blackbear.app hoping to get the latest changes released sometime today -- how do you think they stack up? also how are you testing performance?

(note, you can test the spreadsheets by continuing through the login page at https://my.blackbear.app as an anonymous user)

Seems very cool! I like the concept. I am still designing a "project" sort of organization primitive myself, which would incorporate more time based views of progress.

Just from browsing around and playing with it a bit, it seems there's still a lot of polish needing to be done. Quite a few areas where data is leaking across UI boundaries and some of the hard problems like data management and retention are stubbed (export is nice but it's the floor imo)

Pax is just what I call the agent inside of Blackbear. It's a model agnostic harness, e.g. you can use Claude (Fable, Opus, Sonnet, Haiku), GPT 5.6, or Qwen to provide the inference.

The harness gives the model a set of tools and guidance for working within Blackbear, so Pax regardless of the model you supply can help you write notes, manage a spreadsheet, organize your space, etc.

Oh, and all the editors support real-time collaboration. There's messaging, voice and video calls, and a lightweight contact management system that I'm still iterating on

I left my job as a principal software engineer a few months back and have been working on an agentic harness and https://blackbear.app -- a local first, end-to-end encrypted everything app with file editors for notes, documents, sheets, free-hand canvases, slides, and a whole lot more. Also features a calendar and task system, and a lightweight discovery engine. There's a BYOK agent system (Pax) that works with you across the whole app (OpenAI, Anthropic, Gemini, and local Ollama support). Works in a browser, from the CLI, and from native mac and linux apps (mobile on the way). Not exactly sure my end goal yet, but it's been fascinating building these things