kenhorn.bsky.social
Founder at cloudsinc.net, tabular ML, cloud, fintech, trading, snow, London, CTO, bass.
41 posts
158 followers
329 following
Regular Contributor
Active Commenter
comment in response to
post
BTW, the idea that all this firing of people is about lower taxes for the rich is kind of a smokescreen imho. The real plan is to fork off the functions of the US govt to private people, who buy the functions with credit thus creating a new ruling class of Trump loyalists. A gang.
comment in response to
post
comment in response to
post
3. "Oh, look over there, something arbitrary to distract you!"
comment in response to
post
It'd make sense for projects to rank their tech debt - should factor into dependent product roadmaps. Need visibility to be easy - npm audit, say.
comment in response to
post
Time for a browser extension, I guess now the hard part is done.
comment in response to
post
Agree. There will be many footguns, hence I'd expect simple additions in first release (or even by now, post release).
Are the URLs stable over time/users? (Assume it's a security thing, but shame for tooling)
comment in response to
post
Using AI to clean data is a decent route, but you need a careful approach - classical ML Vs quantum LLM has advantages here. You need to validate and avoid making it worse.
comment in response to
post
For once I have coffee to hand .. Snowman a bit forlorn this morning.
comment in response to
post
Doubling down.
comment in response to
post
Something in your eye.
comment in response to
post
+1 , or max ram, invocation secs, CPU, GPU, disk..
Docker for deployments to evade zip size?
comment in response to
post
I stopped using any sagemaker, too easy to have runaway costs, it needs its own cost dashboard with realtime monthly cost projection overlaying the menu
comment in response to
post
Shurely, buckatabase?
comment in response to
post
search.app?link=https%3...
comment in response to
post
There are many attempts it seems. Eg I use arc.codes which has something. Obvs in a dynamic language you can't cover all reflection etc, but am hoping for help on huge 3rd party libs (eg sklearn).
comment in response to
post
Please nudge Mike, one of the few people I keep an x account for.
comment in response to
post
www.crowdsupply.com/cmg-research... time to revisit?
comment in response to
post
So the prompt is the source, but also the schema of the app produced to allow updates / evolution? LLM for migrations or just import into new schema each re-gen?
comment in response to
post
Agree, that the amazing simplicity of interfacing to an LLM for the power you get is hard to beat. Working on that problem currently.
Big advantage is much less data needed since you are inferring far less structure.
comment in response to
post
I wish @ThreshedThought was here for his Ukraine commentary.
comment in response to
post
That the Google Drive-a-like, IAM thing? Looked at it, but felt complicated for a client I had (but might've been a great fit with enough effort)