Leynos

9 ms·

121.
by Leynos 8mo ago
It's pretty ungainly for anything more than toy models. The rendering's rather hideous too. If you just want pretty C4 models, Draw.io is your best bet. But then you lose all the fun linkage between your code and models.
122.
by Leynos 8mo ago
Trying to sign up with GitHub gives me 404 error.
123.
by Leynos 8mo ago
Brought to you by the same people who think that it should be possible to sack an employee without cause and without notice.
124.
by Leynos 8mo ago
Filling out an application form by hand.
125.
by Leynos 9mo ago
Don't you still need to unit test and version control the SQL artefact that is produced? You need to be able to see which query was used on which date and how it was validated. (Prompts need to be version controlled too, of course)
126.
by Leynos 9mo ago
You don't think Anysphere did?
127.
by Leynos 9mo ago
I guess you could say that is perhaps why the suggestion made by the post being replied to doesn't make much sense to me.
128.
by Leynos 9mo ago
They don't make sense yet. It works out very expensive at present. Far too expensive to run as anything other than an experiment.
129.
by Leynos 9mo ago
Would you be willing to share these links? ROT13 or base64 encode the URLs.
130.
by Leynos 9mo ago
Here's a counterthesis: This is people having fun with a new technology that is far from perfect, is full of unknowns, but is ripe for exploration and discovery. Gas Town itself is a piece of speculative fiction: throwing out a hypothe
131.
by Leynos 9mo ago
Take a look at CodeRabbit and Sourcery if you want to give that a go.
132.
by Leynos 9mo ago
They're leveraging the (relative) ubiquity of npm amongst developers.
133.
by Leynos 9mo ago
Addendum: Tables from the study available here: https://www.pnas.org/action/downloadSupplement?doi=10.1073%2...
134.
by Leynos 9mo ago
Is paying $20 to read this the only option?
135.
by Leynos 9mo ago
I find some of it interesting. I'm very interested in understanding why others' experience of using genAI is so vastly different to my own. (For me it's been as transformational a change as discovering I could do my high scho
136.
by Leynos 9mo ago
I tried GLM 4.7 in Opencode today. In terms of capability and autonomy, it's about on par with Sonnet 3.7. Not terrible for a 10th the price of an Anthropic plan, but not a replacement.
137.
by Leynos 9mo ago
It's all my fault. I make sure I burn through my weekly limit every week.
138.
by Leynos 9mo ago
I pretty much do most of my AI coding in Rust. Although I do still use Python or Typescript where appropriate.
139.
by Leynos 9mo ago
It's the following that is problematic: "I asked each of them to fix the error, specifying that I wanted completed code only, without commentary." GPT-5 has been trained to adhere to instructions more strictly than GPT-4. If
140.
by Leynos 9mo ago
If you're an API customer, you can pin to a specific dated snapshot of the model. See the "Snapshots" section on these pages for GPT-4o and 4.1, for example: https://platform.openai.com/docs/models/g
141.
by Leynos 9mo ago
Codex and Claude Code allows you to opt out of model training. Perhaps you don't believe OpenAI and Anthropic when they say this, but it is a requirement upon which most enterprise contracts are predicated.
142.
by Leynos 9mo ago
It's something of a thought terminating cliché in Hacker News discussions about large language models and agentic coding tools.
143.
by Leynos 9mo ago
Opus 4.5 was released 24th November.
144.
by Leynos 10mo ago
Use review bots (CodeRabbit, Sourcery and Codescene together work for me). This is for my own projects outside of work, of course. I use Terragon for this. 10 parallel rust builds would kill my computer. Got a threadripper on its way throug
145.
by Leynos 10mo ago
pglite might be an option.
146.
by Leynos 10mo ago
It does https://twitter.com/OpenAI/status/1963697012014215181?lang=e...
147.
by Leynos 10mo ago
The AI village was given the goal of spreading acts of kindness: https://theaidigest.org/village/goal/do-random-acts-kindness
148.
by Leynos 10mo ago
While I suspect "hot take" is apt, this wasn't exactly anti-AI. Rather, the author is advocating for their particular way of doing genAI output parsing (as opposed to constrained decoding structured output, they advocate unco
149.
by Leynos 10mo ago
This was why server-side compaction in GPT-5.2 was such a big deal. The model is by default provided with a tool that will prioritise the initial task and salient updates in context window compaction, and the new model has been trained to u
150.
by Leynos 10mo ago
If a model costs them 2x as much, they charge 2x as much. That much is clear from their API pricing.

More ›