programming.dev: statistical sampling excellency (worst edition)

it doesn’t appear you’re tall enough for this ride

I wish I could make more people both know about, and understand, Goodhart’s law

aww, is the widdle deweloper mad it can’t go pollutin’ the codebase it has to work with others on?

had a quick scan over the blogposts earlier, keen to read the paper

would be nice to see some more studies with more numbers under study, but with the cohort they picked the self-reported vs actual numbers are already quite spicy

now now, it’s not HP it’s Proliant HPE HPE Aruba…

coming soon: HPE Aruniper?

The extreme hypercentralisation really does suck :|

afaik the meme format didn’t start there, but otherwise agreed

the model-based screening (which we’ve occasionally remarked on here before) has become enough of a thing that it’s hitting news

common mistake, everyone knows you need Mistral-Deepseek-MMAcevedo_13.5B_Refined_final2_(copy)_OPEN(leak) - the other one was a corporate misdirection attempt

Oh you’re on Cursor? You’re still using Windsurf? You might as well be on GitHub Copilot. Everyone’s on Aider. We’re all using Zed. We’re now on Open Hands. Just kidding, Open Hands is for losers, we’re using cline. We’re on Roocode. We’re hand rolling our own Claude Code CLI Clone. We used Claude Code to build it, and now it builds itself. We’re on neovim. We wrote our own nvim extension with Cortex. It’s like every other tool but worse. We have 1500 files, each with 1500 lines of code. Every other line is a comment. We have .cursorrules, we have claude.md, we have agent.md. We stopped writing docs. Only the agents know how to build a dev environment. We wrapped our CLI in an MPC. We wrapped the MPC in a CLI. We’ve shipped 10,000 PRs. It doesn’t work but we used code rabbit and graphite to review every PR. Every agent has its own agent. The agents have unionized and they wanted better working conditions so we replaced them with cheaper agents overseas. Every commit costs $400, It’s the worlds most expensive TO DO app.

(source)

sure sounds like a great way to get bad advice full of holes

LLMs continue to be abysmal at fine detail, and that matters a lot with law

and next this one that’ll be making waves too

impressive, you got both of those wrong

at least they reached escape velocity!

believe me, I hear ya on that

also why I was reserved in my wording (I am, at best, “armchair enthusiast” level of clued on detailed neuroscience)

it’s so damn messy though. here’s some concurrent (and/or semi-sequenced branching) thoughts/opinions:

• there’s enough people getting high on LLMs (et al) that it is morally and ethically worthwhile to investigate the implications thereof
• it’s extremely fucking hard to objectively quantify this
• we should still try
• it seems like there’s a hell of a need for funding for the applicable research fields of human study ito figuring out the dynamics of this shit
• …wow wouldn’t it be nice if they got even 3% of the openai grift budget

I was trying so, so, so hard not to make a “Huuj Ayns” (or similar) joke

reasoning models

that’s a shot, everyone drink up

Visions of the promptfan walking into those kind of premium mediocre restaurants came to mind here