h14h

12 ms·

61.
by h14h 7mo ago
Probably depends on what is in the merge. Lately I've been collaborating a ton on PRDs and software specs in markdown (now that agents have gotten pretty good at turning it into usable code) and using git had been pretty painful. Espec
62.
by h14h 7mo ago
I am much more interested seeing the M5 Pro and M5 Max in the Mac Mini & Mac Studio. The temptation of running a local LLM on my gaming PC's GPU finally gave me the incentive I needed to set up Tailscale & Mosh, and there'
63.
by h14h 8mo ago
Not surprising, but it's sad to accept that the only major company building consumer-focused computing devices is Apple. My hope is that LLMs allow linux to gain market share quickly. I know personally I've had a much smoother tim
64.
by h14h 8mo ago
Oh dang I had no idea that gpt-oss-120b was that cheap these days. And yeah, given Vultr inference is in beta, their docs ain't great. In addition to kimi-k2-instruct and gpt-oss-120b, they currently offer: deepseek-r1-distill-llama-70
65.
by h14h 8mo ago
Fair point. > It's nothing short of invigorating to have this degree of control over something so powerful I'm a SWE w/ >10 years, and you're right, this part has always been invigorating. I suppose what's &qu
66.
by h14h 8mo ago
I'm using kimi-k2-instruct as the primary model and building out tool calls that use gpt-oss-120b to allow it to opt-in to reasoning capabilities. Using Vultr for the VPS hosting, as well as their inference product which AFAIK is by fa
67.
by h14h 8mo ago
I'm presently in the process of building (read: directing claude/codex to build) my own AI agent from the ground up, and it's been an absolute blast. Building it exactly to my design specs, giving it only the tool calls I nee
68.
by h14h 8mo ago
Yup. Been building my own "Claw" in Go using cloud LLMs and it's running very happily on a $6/mo VPS with 1 vCPU and 1GB of RAM.
69.
by h14h 8mo ago
Having a similar experience. Durable memory with accurate, low-latency recall is not at all easy. Loads of subtle design decisions to make around how exactly you want the thing to work.
70.
by h14h 8mo ago
Is this not just Ralph with extra steps and the risk of context rot?
71.
by h14h 8mo ago
For sure. I imagine it'd be pretty difficult to evaluate the "correct" amount of steer-ability. You'd probably just have to measure a delta in eagerness on a single same task between when given highly-specified prompts,
72.
by h14h 8mo ago
Would be really interesting to see an "Eager McBeaver" bench around this concept. When doing real work, a model's ability to stay within the bounds of a given task has almost become more important than its raw capabilities no
73.
by h14h 8mo ago
OpenAI, Anthropic, and other model providers have created tools (the LLMs) with unprecedented new capabilities. The key problems are a) these new tools have weird limitations that make them hard to deploy effectively, and b) these tools are
74.
by h14h 8mo ago
My guess is that the true impact of this will be difficult to measure for a while. Most "single-person start-ups" will probably not be high-visibility VC-backed, YC affairs, and rather solopreneurs with a handful of niche moonligh
75.
by h14h 8mo ago
I'm fascinated by how the economy is catching up to demand for inference. The vast majority of today's capacity comes from silicon that merely happens to be good at inference, and it's clear that there's a lot of room fo
76.
by h14h 9mo ago
It'd be easy to simply say "skill issue" and dismiss this, but I think it's interesting to look at the possible outcomes here: Option 1: The cost/benefit delta of agentic engineering never improves past net-zero, an
77.
by h14h 9mo ago
In my experience, using AI coding agents need highly specific success criteria, and an easy way to verify its output against that criteria. My biggest successes have come when I take a TDD approach. First I identify a subset of my work into
78.
by h14h 9mo ago
Given how inexorably tied the US economy is to SFH prices, a drop may do more harm than good. Many families have the majority of their wealth tied up in their home, and another significant portion of it (knowingly or not) in the SFH MBS mar
79.
by h14h 9mo ago
It is going to be fascinating to see how the impacts of this will ripple through the housing market. Two areas I'm particularly curious about are a) the financing of new construction, and b) the sale prices of SFHs.
80.
by h14h 9mo ago
The Tailwind Team's Refactoring UI book was a big eye opener for me. I had no idea how many subtle insights are required to create truly effective UX. I think people vastly underestimate just how much work goes into determining the cor
81.
by h14h 10mo ago
It's not overstating it to say I owe my entire SE career to Ruby. Without it's accessible syntax, I don't know that I would have ever managed to overcome the initial "I have no idea what's going on" barrier. Fo
82.
by h14h 10mo ago
> How do you envision you’ll work through all of the captured notes? Do it all at the end of the day? Not entirely sure yet. My current setup involves an Apple shortcut that sends a text blob to my email inbox with the subject "Note
83.
by h14h 10mo ago
I'm sold enough on this form factor to take a flyer on a pre-order. I've been hunting for ways to minimize friction when quickly capturing random thoughts and this is a novel idea that seems to go further than anything else I&#x27
84.
by h14h 10mo ago
I think this calls out a subtle, but significant difference between private and public companies. Public companies as an asset class have to compete with an open market of other investments, so the incentives drive a min-maxing approach to
85.
by h14h 11mo ago
Fully agree. When I said "consumer linux landscape" & "personal computing platform" I was thinking much more broadly than desktop PCs. Admittedly a hypothetical Arm-based Steam Deck or Framework Laptop were at the fo
86.
by h14h 11mo ago
I'm personally rooting for "business decision" over "turning over a new leaf". If FOSS support is motivated by a clear profit motive, then it'll be viewed positively by shareholders and stick around no matter w
87.
by h14h 11mo ago
I hope this is motivated by shrewd decision-making in response to market pressure, as opposed to being strictly a perception thing. While it would be great for Qualcomm to "do the right thing" in supporting FOSS, I feel much more
88.
by h14h 11mo ago
I suspect it has something to do with LA's large footprint. Comparing to where I'm from in Chicago, LA county is over 4x the land area with less than 2x the population: https://www.comparea.org/r122576+r396479 Don
89.
by h14h 11mo ago
I did a lot of parametric 3D CAD work in college as a MechE student, and switched to webdev after graduating. Recently, I had a hobby project that would have involved dusting off my CAD skills, and was bummed to find that there wasn't
90.
by h14h 11mo ago
Should Google be doing more to support ffmpeg? Yes. Should Google stop devoting resources to identifying and reporting security vulnerabilities in ffmpeg? I cannot bring myself to a mindset where my answer to this question is also "yes

More ›