Top all-time · #17889
David Crespo
Personal blog · crespo.business/posts · on HN as dcreprogrammingwebai
David Crespo's personal blog at crespo.business covers programming, web and ai. Since 2025, 4 different Hacker News readers have linked it 16 times, most often “Overengineering PR Create with Jj”. It is still writing: the last post was on 2026-01-21, and you can follow it by feed.
Too few sampled comments (1) to say how often HN recommends it.?
Why HN links it
The one comment cites this blog for an analogy in a thread on Anthropic and Cursor AWS spending, arguing that total AI spend can rise while cost per task falls. The comment does not assess the writing.
Summarised by Claude from a sample of 1 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Web dev + dad in Chicago
Most-linked posts?
- Overengineering PR Create with Jj3 people · 3 mentions · first linked 2025-05-05
- Is the "cost of inference" going up or down?2 people · 6 mentions · first linked 2025-09-03
What Hacker News says?
Suffice it to say this is not the first time Ed Zitron has been egregiously wrong on both analysis and basic facts. It's not even the first time this week. I wrote a post about his insistence that the "cost of inference" is going up. https://crespo.business/posts/cost-of-inference/
Comment linking /posts/cost-of-inference/
on “OpenAI Needs $400B In The Next 12 Months”Submitted Is the "cost of inference" going up or down? as a story · 9 points
HN discussionThis is not a direct response to this piece, but I wrote a short post about the egregious errors Zitron is comfortable pushing in order to make things sound as bad as possible. https://crespo.business/posts/cost-of-inference/
Comment linking /posts/cost-of-inference/
on “OpenAI Is Just Another Boring, Desperate AI Startup”Submitted Overengineering PR Create with Jj as a story · 4 points
HN discussionMy approach in "LLM-only RAG for small corpora" [0] was to mechanically make an outline version of all the documents _without_ an LLM, feed that to an LLM with the prompt to tell which docs are likely relevant, and then feed the entirety of those relevant docs to a second LLM call to answer the prompt. It only works with markdown and asciidoc files, but it's surprisingly solid for, for example, searching a local cop…
Comment linking /posts/llm-only-rag/
on “Show HN: PageIndex – Vectorless RAG”Submitted Overengineering PR Create with Jj as a story · 2 points
HN discussionYes. All the time. I wrote a tool that does it! https://crespo.business/posts/llm-only-rag/ $ rgd ~/repos/jj/docs "how can I write a revset to select the nearest bookmark?" Using full corpus (length: 400,724 < 500,000) # Answer gemini-2.5-flash | $0.03243 | 2.94 s | Tokens: 107643 -> 56 The provided documentation does not include a direct method to select the nearest bookmark using revset syntax. You may be able to…
Comment linking /posts/llm-only-rag/
on “Is the doc bot docs, or not?”Submitted LLM-only RAG for small corpora as a story · 2 points
HN discussionYou're describing increased spending while calling it increased cost. These aren't the same thing. A task that cost me $5 to accomplish with GPT-4 last year might cost $1 with Sonnet today, even though I'm now spending $100/month total on AI instead of $20 because I'm doing 100x more tasks. The cost per task dropped 80%. My spending went up 5x. Both statements are true. Here's an analogy you may understand: https://…
Comment linking /posts/cost-of-inference/
on “How much Anthropic and Cursor spend on Amazon Web Services”This looks fine to use, but considering you can a achieve the same with a git repo and a 20 line bash script to make the symlinks[0], it seems to me precisely the kind of thing beginners should learn to do themselves rather than trust the magic and assume there’s something complicated going on they couldn’t understand. The fact that all of the commands are the same as the underlying git commands is probably better t…
Comment linking /posts/version-your-dotfiles/
on “Show HN: Lnk – Git-native dotfiles manager”Sounds similar to a short script I use to generate branch names with jj, which has the advantage that you don’t have to name a branch until after you’ve already written the code, so the LLM can name it based on the contents of the diff. https://crespo.business/posts/overeng-pr-create-jj/
Comment linking /posts/overeng-pr-create-jj/
on “Git CLI tool for intelligently creating branch names”Check this out: https://crespo.business/posts/dq-its-just-js/ You don't have to use my implementation, you could easily write your own.
Comment linking /posts/dq-its-just-js/
on “A Faster Alternative to Jq”Funny that everyone is linking the tools they wrote for themselves to deal with this problem. I am no exception. I wrote one that just lets you write JavaScript. Imagine my surprise that this extremely naive implementation was faster than jq, even on large files. $ cat package.json | dq 'Object.keys(data).slice(0, 5)' [ "name", "type", "version", "scripts", "dependencies" ] https://crespo.business/posts/dq-its-just-…
Comment linking /posts/dq-its-just-js/
on “A Faster Alternative to Jq”https://crespo.business/
Comment linking /
on “Ask HN: Could you share your personal website here?”It's a response to this: https://shumer.dev/something-big-is-happening The post is silly, but I do not expect Zitron's commentary to be particularly illuminating as he is a charlatan himself. I could point to many examples, but here is a blog post I wrote about one case of him trying very hard to not understand a simple and familiar situation: https://crespo.business/posts/cost-of-inference/ .
Comment linking /posts/cost-of-inference/
on “Something Big Is Coming (Annotated by Ed Zitron) [pdf]”In fact I do! "I know I sound like an asshole, but I’ve got a serious question: what can LLMs do today that they couldn’t a year ago? Agents don’t work. LLMs - read stuff, write stuff, analyze stuff, search for stuff, 'write code' and generate images and video. And in all of these cases, they get things wrong." https://bsky.app/profile/edzitron.com/post/3ma2b2zvpvk2n This is obviously supposed to be a critique, but…
Comment linking /posts/cost-of-inference/
on “Project Vend: Phase Two”