pbowyer

7 ms·

1.
by pbowyer 15d ago
What would be the correct way to do benchmarking/profiling here? I ask as I've started to run some for Vinyl cache to see what performance issues I can find, and I've quickly learned just how hard it is to get good, repeatabl
2.
by pbowyer 1mo ago
I think we're both looking for the same improvements and I haven't found them in a writing tool yet: https://news.ycombinator.com/item?id=49331036
3.
by pbowyer 1mo ago
I like seeing what people create, and you've created something with some interesting features, particularly around the path map and keeping all drafts. I'm forever duplicating a paragraph and rewording it 6 or 7 times, then deleti
4.
by pbowyer 1mo ago
All of this, but underlying it I've come to the conclusion that Markdown isn't a good format for a writing environment beyond just simple prose. There's too much left up to the flavour of it and it's got weaknesses. I&#x
5.
by pbowyer 2mo ago
Friendly reminder that GPT-5.6 in Codex cannot spawn Luna subagents, only Terra [1] To fix this you currently need to make your own copy of the bundled model catalog [2] and opt Luna into MultiAgent V2. 1. https://github.com&#x2F
6.
by pbowyer 2mo ago
Can you say more? This sounds like it could be more useful than trying to build a knowledge base (say qmd) and then get the models to search it for the right context each time.
7.
by pbowyer 3mo ago
> I feel like this benchmark reiterates my disbelief that anyone uses the latest Anthropic models for any productive work. They seem to be the best at burning tokens and spawning unnecessary subagents even for well-defined and tightly sc
8.
by pbowyer 3mo ago
> I'm only using Opus at all because GPT-5.5 in the subscriptions only has a small (400k, but 258k effective) context window. Do you find that makes a difference in your work? I've been using 5.5 high/xhigh to optimize and
9.
by pbowyer 3mo ago
The one positive of this is it seems to have livened up the open source project (Vinyl Cache). Features that haven't been added for years are now being worked on, whether they step on the toes of the commercial offering or not.
10.
by pbowyer 4mo ago
Jira is the one product I feel needs to be AI native. AI native in the sense that it papers over the pain points. New JIRA admin? AI will set it up to do what you want (after all, Atlassian has a great training set as they can see which Clo
11.
by pbowyer 4mo ago
Previously covered at https://news.ycombinator.com/item?id=48053623
12.
by pbowyer 4mo ago
What's led to the higher prices than Germany? Usually substantially lower earnings would mean lower prices, even if not substantially lower (look at the UK, higher prices than much (all?) of Europe, average earnings slightly less).
13.
by pbowyer 4mo ago
> Extremely debatable. They still have never fully implemented health checks and auto healing. Agree. Plus there's the monitoring of the host that is always overlooked in articles. I've ended up chucking Monit on there to monit
14.
by pbowyer 5mo ago
Only at the "hmm that seems an interesting idea" level. Thanks for the links, going to have a read and see if I can apply any to my work.
15.
by pbowyer 5mo ago
If you can type code fast and accurately, it sounds a great process to use. You're using LLMs for the bit where they bring great value, and yourself as a higher quality coding agent :)
16.
by pbowyer 5mo ago
Developer of 20+ years here, can't give you an accurate multiplier but I am faster. Because spotting holes in specs has never been one of my strengths. And working without technical colleagues much of the time, it's a boon to be a
17.
Lessons from Debugging GLM-5 at Scale(z.ai)
1 points by pbowyer 5mo ago|0 comments
18.
Scaling Pain of Coding Agent Serving: Lessons from Debugging GLM-5 at Scale(z.ai)
12 points by pbowyer 5mo ago|0 comments
19.
by pbowyer 5mo ago
> Hermes is one of these OpenClaw clones So that's what it is. Reading its README I thought it was another harness like Pi [1], but with built-in memory so it remembers what it learns, and gets more capable the longer it runs. Like
20.
by pbowyer 5mo ago
What's the privacy/data security like? I can't find that on that page. Edit: found it. > We may use your Content to operate, maintain, improve, and develop the Services, to comply with legal obligations, to enforce our pol
21.
by pbowyer 5mo ago
> Stripe APIs being simple and easy is a meme from the 2010s. It isn't anymore. I'm working with Stripe subscriptions at the moment for a charity taking donations via their website. The subtle differences between subscriptions
22.
by pbowyer 5mo ago
The Peter Principle: https://en.wikipedia.org/wiki/Peter_principle
23.
by pbowyer 5mo ago
Let us know if it does, because we all want it to work :)
24.
by pbowyer 6mo ago
Dealing with Google is a nightmare. I'm one of the volunteer sysadmins for https://forum.buildhub.org.uk/ , a DIY and self-build forum. For 10 years it ranked very well on Google, particularly in the UK, and then on 28
25.
by pbowyer 6mo ago
This nono? https://github.com/always-further/nono > Just remember to disable CC’s auto-updater if that’s what you’re using. Why?
26.
by pbowyer 6mo ago
Can anyone enlighten me how having a coding harness when for most customers you say "we won't train on your code" helps you do RL? What's the data that they rely on? Is it the prompts and their responses?
27.
by pbowyer 6mo ago
If you were creating a new programming language in 2026, which DateTime/Temporal library would you copy and why?
28.
by pbowyer 7mo ago
Is there only a cloud-based mode for Islo, or can I run this entirely on my laptop?
29.
by pbowyer 7mo ago
I came across a coding harness using Lua as its control plane yesterday: https://github.com/hsaliak/std_slop/blob/main/docs/lua_integ... > std::slop is a persistent, SQLite-driven C++ CLI agent.
30.
by pbowyer 7mo ago
For others, here's the paper: https://arxiv.org/abs/2507.00002

More ›