Pinned Tweet
I've been building Tribal, a memory system that stores tacit engineering knowledge as you work and makes it queryable across your team.
My taste for hyper-configurable systems comes from a previous job, where bank operators configured the system in Python themselves.
I brought that instinct to my own project; when sketching the first config schema, I reflexively made the embedding model as configurable as the inference model. In a config file, they look like the same kind of setting.
They are not the same thing.
Swap your inference model and nothing you've already stored notices. Swap your embedding model and none of it can be searched until every last record is rebuilt.
The new post is about that gap: how a small promise about a system's behaviour can cost far more to keep than it did to make.
Keeping this one meant implementing zero-downtime migration - rebuilding everything underneath a live system without anyone feeling it shift
Follow for more articles like this one
𝐗𝐕 retweeted
wrote about my experience building an eval harness in rust to test a memory system
type-2 fun, learned a tonne, and now I have an eval harness
𝐗𝐕 retweeted
Most agree that, in the near future, agents owned by different systems will need to work together toward higher-order goals
I find this idea incredibly interesting, so I spent Sunday evening exploring where it breaks
I ended up visualising it as a pixelated world starring Ember, Moss and Lumen:
𝐗𝐕 retweeted
Today, we are introducing Inkling.
Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available.
thinkingmachines.ai/news/int…
Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵
It occurred to me I could lean on good old code quality techniques here, so eventually I added this:
"The urge to write or lengthen a comment is a signal to try the code first: rename the value so the meaning rides the name, extract the block into a function whose name carries the why, lift the invariant into a type or an assertion.
"Prose is the fallback for what code structurally cannot say: the outside constraint, the non-obvious why, never a patch over a name that failed to say it.
"A comment that grows under review is the sharpest case: each round's added explanation is a fact looking for a home in the code, and the comment should shrink as the code absorbs it, never lengthen to defend it."
Something has definitely been lost with this approach to building
Not too long ago, I would be happy that I had a new idea to work on
I'd think "finally, something open-ended I can sink my teeth into"
These days, everything feels far more urgent
Racing toward something, whilst running away from something else, not being able to name either
I think it's insane how it's becoming normal to spend over $1000 a day on API calls to generate code.
It compresses time, sure, and the time (days of work, months of study) alone is potentially worth more than the $1000...
...but I haven't adjusted to the economics of any of this at all.
