HN user

noveltyaccount

1,168 karma
Posts1
Comments338
View on HN

I really wish Apple would fund a small team to open source some documentation and drivers to help Asahi along. I know they won't, but I can dream. It would be a drop in the bucket for Apple but would cement their hardware as de facto for silicon valley engineers (even more so than today).

I agree a block list won't work. And unix file permissions may not be enough; I once saw Codex 5.4 use docker to execute a command as root since it couldn't run sudo. Running in a container may be the only solution:

sudo needs an interactive password here, so I'll use Docker itself to prepare the bind-mount directory as root and hand ownership back to UID/GID 1000. That keeps the compose file's non-root runtime intact.

Ran `docker run --rm -v /shares:/shares alpine:3.20 sh -c 'mkdir -p /shares/local-llm/models && chown 1000:1000 /shar...`

I prompted Codex 5.5 to one shot something where I wanted the design to have a pluggable decision module. I gave it a few examples of the kinds of inputs and actions I expected. I did not constrain it beyond that high level of what I wanted. The design it came up with was very good. Easily on par with what any senior engineer at big tech would. And cleanly decoupled in a way that would make future refactoring simple. I was damn impressed.

I have read countless self help books where I complained about fluffy filler material and the book should have been fifty pages instead of two hundred. Not a surprise to me that a topic distilled through an LLM is a better experience.

From the recent Nvidia & Microsoft announcement about new chips for consumers:

“Our goal is to deliver unmetered intelligence to every home and every desk with Windows,” said Satya Nadella, chairman and CEO of Microsoft. “RTX Spark marks a real breakthrough towards that vision.”

Makes me optimistic that those two companies are going to keep investing in quality local models.

I do this with Codex 5.5 for planning (specs, technical design, and task list); and Qwen 3.5-35B for task by task build out. It requires more hand holding and makes more mistakes than using Codex for everything, but it helps me spread my $20 chatGPT subscription pretty far.

“Our goal is to deliver unmetered intelligence to every home and every desk with Windows,” said Satya Nadella, chairman and CEO of Microsoft. “RTX Spark marks a real breakthrough towards that vision.”

I expect computers with this chip will be about $4000. If Microsoft can deliver on local AI models that can orchestrate Windows and have solid real world intelligence, that will be an inexpensive business purchase compared to pay as you go tokens. I'm excited to see how this plays out.

Same. I wish there was a good story for .NET UI on desktop, mobile, and web. I find Maui Blazor Hybrid to be pretty close. The desktop and mobile variants compile to a webview bound to native code (no WASM, no JS), and the Web variant compiled to WASM. Much leaner than Electron. Pretty clunky though last time I tried though that was three years ago.

I'm someone with 20 years in software and the last 10 in management. I have good instincts, good design pattern knowledge, and understand system design well. But my actual coding skills are rusty, I can do it but it takes a lot of time to RTFM because specific libraries and syntax aren't on the top of my mind.

With AI I can build. I'm having so much fun turning ideas into code. I can do a week's worth of work before lunch. I can ask AI to add comments so detailed that my code becomes a refresher tutorial.

It's so exciting to be able to bring my ideas to life, make use of my experience, and not be hobbled by my somewhat atrophied hands-on coding skills. I for one welcome this revolution.

I go for walks or runs--without headphones. If I listen to music or a podcast, my mind doesn't wander. I need that quiet for processing.

I'm using it on a project now and have been very impressed. It really lives up to its goal of getting out of the way and automating the tedium. I think about my Postgresql schema, row level security, and let the magic give me a GQL API with websockets too.

Qwen3-Coder-Next 6 months ago

separate models for /plan and /build

I had not considered that, seems like a great solution for local models that may be more resource-constrained.

Qwen3-Coder-Next 6 months ago

I think I like coding models that know a lot about the world. They can disambiguate my requirements and build better products.

Patterns.dev 7 months ago

In my senior year of college two decades ago, I needed one or two credit hours to finish up, and I signed up for a once per week software patents (as in, intellectual property, I thought) course. It turned out to be a patterns course taught by none other than Ralph Johnson and the text was his famous Gang of Four Design Patterns book. Happy accident, it turned out to be among the most professionally useful courses I ever took.

https://en.wikipedia.org/wiki/Ralph_Johnson_(computer_scient...

I'm with you, a little pogo-pin connector with all the charging circuitry in the external dongle would add, I don't know, ten dollars to the BOM. Very cool product but I won't buy a gadget I know I'd dispose in two years.

In Copenhagen, the vast majority of people park their bikes on the street using only a cafe lock (frame mounted, immobilizes the rear wheel). The bikes are generally nothing special, old rusty junkers, with one or three gears. E-bikes flatten terrain but also you need an indoor place to store it and they become a magnet for theft. A cheap bike you can ride to the Metro and leave in the elements is versatile in a way e-bikes are not. (I say all this as a massive e-bike fan living in a very hilly US city who recently visited Copenhagen and adored its bike culture.)