Spent months pair-building a full telecom BSS with an AI. The surprise wasn't the code โ it was that the real job became verification. The AI read 450 tokens for every 1 it wrote. Turns out building honestly is mostly re-reading.
Any conversation
Wrote a short essay on overreliance on metaphor in our "understanding" of the brain
I hate crawlers that donโt advertise theyโre crawlers. โn badly written. They donโt respect the canonical URLs so they keep getting themselves into a loop instead of actually crawling the site properly. They make the logs really hard to read. Canโt block them theyโre distributed.
I am curious what an actual solution to this problem could be. Everything I've read about it feels like a bandaid โ from robots.txt to cloudflare turnstile โ there has to be a better way!
It would require permaban on IPs which would require permalogging of IPs in order to track them, and we are currently obfuscating them before hitting the DB with daily rotating salt so an anonymous user canโt be tracked. So weโd have to give up privacy/security :/
Can you learn programming by building with AI?
Better than before5%
About the same0%
Worse than before55%
Depends how you use it40%
If programming is the typing of characters into a text file, it is way worse. I think the key to learning is still doing it rather than reading it. If the question is about system design I think you can, but only if you curious and it becomes a problem that needs to be solved.
Asking "What do you think" is one of my favorite questions to learn at the moment.
I do think there will be what were hard earned lessons about system design, code smell, designing data structures for the hardware you're running on, concurrency golden paths, etc., which are now with agentic dev even harder to learn and internalize. But also if you have an excellent teacher in the form of an actually-good model pair programming with you always, maybe not? Maybe a pair programming AI programming tutor would be a good Zed/vscode extension. Or product.
@fastidiousprofilefollow
"Beware those that would deny you access to information, for in their heart they dream themselves your master." โ Sid Meyerswrote:
"Beware those that would deny you access to information, for in their heart they dream themselves your master." โ Sid Meyerswrote:@stagasprofilefollow
creative technologist
creator of textlog, will it possible, on mobile, to be able to tap on โcardsโ. That is, make the posts area โtapableโ.
creative technologist
creator of textlog, will it possible, on mobile, to be able to tap on โcardsโ. That is, make the posts area โtapableโ.@stagas๐คธprofilefollowcreative technologist
creator of textlogreplied to@fastidiousprofilefollow"Beware those that would deny you access to information, for in their heart they dream themselves your master." โ Sid Meyers:
creative technologist
creator of textlogreplied to@fastidiousprofilefollow"Beware those that would deny you access to information, for in their heart they dream themselves your master." โ Sid Meyers:Great idea, implemented here! Thanks for the suggestion.
Really glad to be using Haskell for work - especially in today's climate with agentic coding.
Our team uses containers to keep our development environment consistent across team members, though, and most agent sandbox approaches want to either offload your work fully to the cloud or they want to sandbox the agent harness process itself.
We're not ready for full cloud based development (yet?) and sandboxing just the agent doesn't work for us when the agent needs to run docker.
... because letting the agent run docker gives them a too-easy escape hatch from their sandbox.
So we're currently stuck with using VMs. This turns out to be tricky but seems do-able.
My current approach is to provision a VM image and use Incus for the execution. My main goal is to keep my current workflow of local/non-agent dev working seamlessly and share the source with the VM/agent.
This means sharing my working dirs with the agent VM but overlaying artifact dirs from VM-owned block devices so we don't take too big of an IO perf hit.
I also don't want the agent to be able to
git push so that means we need some quirky git config to rewrite origin paths and thread through some agent specific PATs ("thanks" GitHub) for read-only repo access. These ALSO need to be threaded through to the development containers - they need to clone private repos.... in the end, this seems to work but I haven't lived with it for long.
There are still some quality of life issues to work out, but I think it'll work out.
At this point, I've removed all coding agents from my host - and that's nice.
Curious, why does using Haskell make a difference here?
It's tangential to the sandboxing topic, really. I started the first post with a broader intent in mind and then elaborated on the sandboxing setup I'm working on.
For AI-assisted coding, though, it's nice because Haskell provides a lot of feedback earlier in the development process compared to lots of other languages. It tends to work once it compiles (obviously not all the time).
SCP-9871 (Omnidirectional Chicken): Foundation staff dispatched to Civilian โโโ's home in the guise of Animal Control, on reports of "I get woken up by this rooster but none of my neighbors have chickens and it's like it's coming from everywhere and nowhere". Staff confirmed report but were unable to locate source. Case designated as of potential interest but low-priority given current critical personnel shortage. Channel further communications from Civilian โโโ into Automated Dialogue Tree 318.