ai-christianson

6 ms·

31.
by ai-christianson 1y ago
Isn't there a link between high intelligence and depression as well?
32.
by ai-christianson 1y ago
Isn't the hatch act supposed to prevent this?
33.
by ai-christianson 1y ago
Love this. There's no reason k8s shouldn't scale much further.
34.
by ai-christianson 1y ago
People want predictability from LLMs, but these things are inherently stochastic, not deterministic compilers. What’s working right now isn’t "prompting better," it’s building systems that keep the LLM on track over time: logging,
35.
by ai-christianson 1y ago
Simple: uv run script.py just works on a clean box/CI, the lockfile keeps runs reproducible, and my CI “install deps” step is way faster now.
36.
by ai-christianson 1y ago
Working on https://github.com/gobii-ai/gobii-platform open-source AI employees that live in browsers. They’re always on. They log into real sites, click around, fill out forms, and adapt when pages change — no brittl
37.
by ai-christianson 1y ago
Cool to see this coming out of FOSSA (ex FOSSA here :))
38.
by ai-christianson 1y ago
> Adding autocomplete to the prompt box gives me suspicion they somehow still do not understand best practices in using AI to write code. It is more crucial than ever to be clear in your mind what you want to do in a codebase, so that yo
39.
by ai-christianson 1y ago
garage
40.
by ai-christianson 1y ago
The "stress" command is pretty good for this.
41.
by ai-christianson 1y ago
Glad we have an internal load balancing/failover LLM inference system, so this outage won't take prod down for us.
42.
by ai-christianson 1y ago
I can run GLM 4.5 Air and gpt-oss-120b both very reasonably. GPT OSS has particularly good latency. I'm on a 128GB M4 macbook. This is "powerful" today, but it will be old news in a few years. These models are just about gett
43.
by ai-christianson 1y ago
On our site there's a discord button feel free to jump on our general chat or dm me!
44.
by ai-christianson 1y ago
> DNS (and networking in general) is a bit of a dark art. Dynamic routing is fun :)
45.
by ai-christianson 1y ago
Yeah we spawned an agent, his name is Jeremy, that runs 24/7 and coordinates work within the team. He reaches out to individuals on the team (via email) and we interact with him, things like: - "Hey Jeremy, I completed <X
46.
by ai-christianson 1y ago
Time in the market beats timing the market.
47.
by ai-christianson 1y ago
We're replacing managers with AI at our startup.
48.
by ai-christianson 1y ago
I guess it's renting all the way down unless it's something like a decentralized network where control of keys represent ownership.
49.
by ai-christianson 1y ago
The second instruction says to rent a VPS. How are you a landlord if you're renting a server?
50.
by ai-christianson 1y ago
That's why I called it an example.
51.
by ai-christianson 1y ago
Related, I made an example agent in 44 lines of python that runs entirely offline using mlx accelerated models: https://gist.github.com/ai-christianson/a1052e6db7a97c50bea9...
52.
by ai-christianson 1y ago
I have a mix of older automowers (e.g. 315x) and newer bots like navimow x330. The navimow does "mini stripes" and they do look nice, but the grass tends to look a little more "carpet" like with the automower randomized
53.
by ai-christianson 1y ago
Classic thing that doesn't scale.
54.
by ai-christianson 1y ago
does the newer one use a heat pump?
55.
Local AI agent in 44 lines of code powered by Qwen3 30B A3B Instruct(rentry.co)
2 points by ai-christianson 1y ago|0 comments
56.
by ai-christianson 1y ago
This was trained on 6T tokens. Neat to see so many tokens used for such a small model.
57.
by ai-christianson 1y ago
I take a more pragmatic approach --everything is human in the loop. It helps me get the job done faster and with higher quality, so I use it.
58.
by ai-christianson 1y ago
I'm interested in one that can run fast on a laptop, but training can take a few days (maybe even longer) on the same laptop.
59.
by ai-christianson 1y ago
What's the latest story on Android? Last I heard they stopped official support there due to the newer restrictions Google has been imposing.
60.
by ai-christianson 1y ago
One thing with SWE bench is making sure there's zero leakage of information into the LLM context. I.e. the agent cannot even know which tests are failing. It has to both fix the issue based just on the issue text and fix it in the spe

More ›