My sentiments exactly. All the formats were so poorly read that they are all effectively useless.

The momentum behind vibe coding is driven by an underestimation/misunderstanding of what it takes to build software.

So much about this seems inauthentic. The post itself. The experience. The content produced. I wouldn’t like to be on the other end of the production of this content.

This just sounds like a normal day for someone who does research and analysis in 2025.

Where do you think expert analysis comes from?

Talk to experts, gather data, synthesize, output. Researchers have been doing this for a long time. There's a lot of grunt work LLM's can really help with, like writing scripts to collect data from webpages.

Great! You’re not the audience for it.

Who is?

The people who pay for what I do.

What happens on the day when those people just directly pay some AI model to do it?

Then that part of my work will change.

However, as this thread demonstrates repeatedly, using LLMs effectively is about knowing what questions to ask, and what to put into the LLM alongside the questions.

The people who pay me to do what I do could do it themselves, but they choose to pay me to do it for them because I have knowledge they don’t have, I can join the dots between things that they can’t, and I have access to people they don’t have access to.

AI won’t change any of that - but it allows me to do a lot more work a lot more quickly, with more impact.

So yeah, at the point that there’s an AI model that can find and select the relevant datasets, and can tell the user what questions to ask - when often they don’t know the questions they need to have answered, then yes, I’ll be out of a job.

But more likely I’ll have built that tool for my particular niche. Which is more and more what I’m doing.

AI gives me the agency to rapidly test and prototype ideas and double down on the things that work really well, and refine the things that don’t work so brilliantly.

So is using an LLM to write SQL for you like using a mitre saw instead of a table saw? I guess the crux is that you still need to do work either way.

Using LLMs to write SQL is a fascinating case because there are so many traps you could fall into that aren't really the fault of the LLM.

My favorite example: you ask the LLM for "most recent restaurant opened in California", give it a schema and it tries "select * from restaurants where state = 'California' order by open_date desc" - but that returns 0 results, because it turns out the state column uses two-letter state abbreviations like CA instead.

There are tricks that can help here - I've tried sending the LLM an example row from each table, or you can set up a proper loop where the LLM gets to see the results and iterate on them - but it reflects the fact that interacting with databases can easily go wrong no matter how "smart" the model you are using is.

> that returns 0 results, because it turns out the state column uses two-letter state abbreviations like CA instead.

As you’ve identified, rather than just giving it the schema you give it the schema and a some data when you tell it what you want.

A human might make exactly the same error - based on misassumption - and would then look at the data to see why it was failing.

If we assume that a LLM would magically realise that when you ask it to find something based on an identifier which you tell it is ‘California’ it would magically assume that the query should be based on ‘CA’ rather than what you told it, then that’s not really the fault of the LLM.

I expect one of the database MCPs would do the trick too, if you didn't mind burning _several_ rounds of back-and-forth under the covers.

We have Woom bikes. This is exactly what they recommend. Each of my three kids learned at different speeds, though. It basically turns your pedal bike into a strider bike.

I lived there between 2004 and 2023 and used to walk along the perimeter of the protected dunes. I didn’t know it was an intentional engineering. I thought it was just natural dunes that were protected. They are really beautiful and you get a sense that there’s a lot of life going on. I always wondered why Santa Monica’s beaches were so barren otherwise compared to other beaches on both the West and East coasts. Good read!

This usually means the military.

What a great example of an https://en.wikipedia.org/wiki/Aptronym . This guy's last name is Heersink and he's finding out where all the boats sank.

Reminds me of Andrej Karpathy. "car pathy" And he worked on Tesla Autopilot.

And Nominative Determinism https://en.m.wikipedia.org/wiki/Nominative_determinism

Wow that article is full of gold: * Justice named Igor Judge * Lawyer named James Counsell * Lawyer named Sue Yoo * Weather reporter named Storm Field * Vicar named Michael Vickers * Orthopedic surgeons named Limb, Limb, Limb, and Limb

Sometimes I feel like the universe really is a simulation and these names are the result of QA testers having a little bit of fun keeping track of who's who.

Interesting thing is he spent eight years creating a comprehensive dataset of over 14,500 mapped records of sunken ships...why??

Going on 12 years now. If I had known it would have taken so long before I started I might not have taken this on. It’s been a lot of work on my own time but it’s something I enjoy. Why did I start? I thought it would be cool to see it all mapped and see the patterns that result, including the temporal ones. And I think I was right - it is cool. Also, aside from this forum there’s been a huge interest in from folks involved with cleaning up potentially polluting wrecks and underwater cultural sites. Glad people are enjoying it and finding it worthwhile.

I applaud such hobbies. Compare the time spent on it relative to, say, ingesting Netflix into your optic nerve.

Yeah, you're right I guess. Better than popping your eye watching cringe shows.

People have obsessive hobbies. There are tons of examples.

This is disappointing. I thought Boar’s Head was the high quality stuff.

I had never heard of them before yesterday when I happened to be browsing On a Roll's menu touting their Boar's Head BLT which I mistook for an actual cut of meat. So it was a selling point before.

I was a CS undergrad at MIT and took the fancy math people’s analysis class for a math requirement and they used Rudin and it killed me. We went through like 170 pages of it for one semester. The professor was Sigurdur Helgason. I went into office hours one day and ask him a question. He slowly walked to his window and replied “the ravines in Iceland are deep” and it was at that point that I realized I was an engineer and not a mathematician.

Lol. Did you figure out what that was supposed to mean?

Not the OP but as a mathematician I can say this, if you study analysis and start to ask questions about why or how certain things work, you will quickly fall deeper and deeper into a rabbit hole until you reach the basic axioms. In analysis the distance between these axioms and what people use analysis for on a daily basis can be quite large.