mafro

I challenge you to use a frontier LLM to analyse your code and specs and verify if they are meaningfully aligned.

Code almost never speaks to the WHY; ADRs will help there.

Code never captures requirements, but it does reflect actual system behaviour, which is a specification.

I can appreciate your approach, but I'm not hand writing 100s to 1000s of specs by hand - at that point I'll just write the code myself

I've used OpenSpec extensively on two large pieces of software which were worked solo over 6-9 months.

Recently, I've completely ditched the specification part. I found they just weren't useful over the longer term. I used an LLM to assess in both directions whether the code matched the specs and whether the specs matched the code. On both software projects this came out with huge divergence from spec to code.

Basically the old theory is true - the code IS the specification.

What I did find very useful and have retained is the process flow. Create a proposal, review the proposal, implement, review the code. Also useful was building and maintaining ADRs and invariant logs for where a unit test cannot be made to verify behaviour. The process and the ADRs, unit test, invariant log all help the software stay coherent as the LLM churns on it over many unconnected contexts.

The author doesn't distinguish whether these projects failed because of technology and execution, or failed because product market-fit. They simply blame AI being involved.

AI will only help if you use rapid iteration to cheaply/quickly produce ideas. All the normal project failure modes still exist.. Blaming AI because AI is dumb.

I only read up to point 3 because the hyperbole and frothing fervour was overwhelming.

I didn't know what BGP is, but I did know CTF. YMMV

Never let a good crisis go to waste

This comment feels like a troll.

I'm sure you are aware that:

- Most open source projects on the Internet are on Github. One could argue Github itself has taken a significant role in the growth of OSS in the last decade

- Using Discord is totally du jour for open source communities in 2025

- Reddit is still the biggest discussion forum on the Internet. Including discussion about open source

There are only two essential apps on my Mac. iTerm2, and a browser

The same idea applies to self-driving cars though, no? That is an industry where the "AI revolution" will enrich only the existing incumbents, and there is a huge bar to entry.

Self-driving cars are not going to create generational wealth through invention like microprocessors did.

Latest release Apr 2025 introduced a plain text save file format, which plays nicer for source control. Great to see development is still active.

Ask an LLM to transcribe and give the overview and key points

This is a precursor to a future model which isn't simply a router.

From the system card:

"In the near future, we plan to integrate these capabilities into a single model."

One reason for this release is surely to respond to their mess of product line-up naming.

How many people are going to understand (or remember) the difference between:

GPT-4o GPT-4.1 o3 o4 ....

Anthropic and Google have a much better named product for the market

One of the few Facebook groups I stayed in over time. It has a very British sense of self-deprecating humour. We're all amused by our mutual dullness.

Great post. I'm fairly new to the AI pair programming thing (I've been using Aider), but with 20 years of coding behind me I can see where things are going. You're dead right in the conclusion about now being the time to adopt this stuff as part of your flow -- if you haven't already.

And regarding the HN post getting buried for a while there...[1] Somewhat ironic that an article about using AI to help write code would get canned for using an AI to help write it :D

[1]: https://news.ycombinator.com/item?id=44214437

Thanks for this breakdown, I guess I'm largely in the window of points 3-6.

Any suggestion on where to start with point 1? (Also a SWE).

That's the whole point of /architect mode, no? You refine the solution in the prompt before aider asks you if you want to apply the changes.

I switched to hatch last year for many projects, its been quite pleasant.

Has anyone used both hatch and uv, and could comment on that comparison?

EDIT: quick google gives me these opinions[1]

[1]: https://www.reddit.com/r/Python/comments/1gaz3tm/hatch_or_uv...

I've been using tailscale/tailscale-caddy[1] successfully to serve applications only on my tailnet. It says highly experimental, but it's worked just fine for me.

[1] https://github.com/tailscale/caddy-tailscale

Just noting that this narrative in the post is contrary to the author's own guidelines - where they go into not finishing books (called "DNF"s lol)

https://tracydurnell.com/2023/08/19/my-reading-philosophy-an...

In this pricing model, the power users effectively subsidise the entry-level users.

Not saying I agree with it, but Apple has been doing this successfully for many years.

Currently there's only 9 cities supported worldwide, and of those 2 are in beta. If this software had broader reach, it may be easier to get open source support.

Having said that, I'd add my city if it were straightforward. It looks like you've spent a lot of engineering time in library and SDK support lately - I suggest investing in the DX happy path to encourage new folks to invest their time.

For example, this document is rather complicated https://github.com/OneBusAway/onebusaway/wiki/Multi-Region

Played through with my 7yo daughter. She loved the game, but would freak out when chased by the "rats". It was a good opportunity to show her how to try to keep playing when you're stressed! She found it exciting after a couple of goes. You didn't mention how old your daughter is, but that is surely the key factor.

I'm still hopeful for a "Stray 2" - or whatever they can do now the devs lost the IP. My daughter cried when at the game's ending (no spoiler)

I tried shairport some time ago, but settled on https://github.com/badaix/snapcast and have a really great and reliable open-source multi-room setup

Why would one snapshot help?

Out of interest, what other tools might be "better" than Pandas?

I like pandas, and python.

I happened to try out Hatch this past weekend, and once I grokked the conventions it largely just worked.

If you like the look of Rye, I suggest trying Hatch also.

Amazing! Any idea when the third edition will be published?

Thanks for your work on this

Could you share the data for this please?

Not disagreeing with your statements, but the key question in the research is why did depression spike in children worldwide _after 2012_.

At various points in my lifetime there have been reasons to be sad about the world - but what changed in 2012 that predominantly appears to affect teenage girls, and has been observed in several countries globally?