AskHN Digest

1. Ask HN: Share your AI prompt that stumps every model

Top comment by thatjoeoverthr

"Tell me about the Marathon crater."

This works against _the LLM proper,_ but not against chat applications with integrated search. For ChatGPT, you can write, "Without looking it up, tell me about the Marathon crater."

This tests self awareness. A two-year-old will answer it correctly, as will the dumbest person you know. The correct answer is "I don't know".

This works because:

1. Training sets consist of knowledge we have, and not of knowledge we don't have.

2. Commitment bias. Complaint chat models will be trained to start with "Certainly! The Marathon Crater is a geological formation", or something like that, and from there, the next most probable tokens are going to be "in Greece", "on Mars" or whatever. At this point, all tokens that are probable are also incorrect.

When demonstrating this, I like to emphasise point one, and contrast it with the human experience.

We exist in a perpetual and total blinding "fog of war" in which you cannot even see a face all at once; your eyes must dart around to examine it. Human experience is structured around _acquiring_ and _forgoing_ information, rather than _having_ information.

2. Ask HN: My CEO wants to go hard on AI. What do I do?

Top comment by codingdave

From what I've seen, anyone giving you definitive answers is blowing smoke. I've seen teams go for AI, do it well, and succeed. I've also seen teams go for AI, f it up, and all get fired as the company shuts down.

The question is whether the key value of your product can benefit from the strengths of AI? If not, don't go there. If so, you then need to determine if your team can actually deliver an AI-driven vision that enhances the existing value prop. Again, if not, don't go there. If so, do it.

But from your description, your CEO is not asking those questions - they are asking, "How do we get more funding?" Which tells me that your CEO doesn't give a crap about building a product, they are just trying to make money and get some nice bullet points on their resume about the size of a company they led.

That puts you in the position of choosing whether you want to go on a VC-driven startup ride just to have the experience, or whether you want a product-driven role. People have their reasons for both directions, but if you want a product-driven role, you are out of alignment with your CEO and probably shouldn't work for them.

3. Ask HN: How do you talk about past jobs you regret in interviews?

Top comment by futureshock

An interview is a sales pitch for a product. The product just happens to be you. Set aside whatever negative feelings you have about this previous job or the people you worked with there. The interviewers care if you will do their job well and with consistency and professionalism. Your personal feelings are irrelevant as long as you can keep them to yourself, or maybe tell your dog.

ANY negativity during a job interview is going to work against you. It is expected that you find a way to spin every situation and every project in some kind of positive light. Even when interviewers ask for weaknesses or about conflict, the “right” answer is to be able to talk about that negative thing in a way that lets your true brilliance shine through. Skilled candidates know how to inject just the right amount of humanity and relatability in an otherwise perfect employee.

If you are having trouble separating your feelings from your ability to keep to your talking points, then a good therapist may be able to help you learn better emotional regulation skills.

In the future, keep working to proactively manage your career. Keep yourself in roles where you are learning and thriving. When you feel burnout creeping in, deploy strategies to counter it or at least get yourself into a new situation.

4. Ask HN: Why do we celebrate AI-Copilots but reject AI–Generated art?

Top comment by Sindisil

Who is "we", Kimosabe?

Personally, I have no time for gen-AI in pretty much any context, at least given the current landscape.

And plenty of people seem to accept, if not love, gen-AI art. I don't get it, but it's true.

> While browsing YouTube, an AI-generated video appeared and I reflexively told my wife, “That’s AI—skip it.”

My reflex whenever I encounter gen-AI output in any form: text, code, image, music, video, what have you. I find all off it mid in the best of cases, and usually think it's quite terrible. I regularly see posts of the form "you'll never believe this amazing AI generated picture/video/paper/program, and when I check it out I feel like I'm taking crazy pills because I just don't see the magic.

Just my $.02, not inflation adjusted. You (and many others) may well feel differently.

5. Ask HN: What did you learn too late in your tech career?

Top comment by proc0

That ultimately, the industry wants engineers to manage the business, instead of managing software.

If I had known this I would not have gone through university at all. I would have attempted my own business from the start and learned on my own.

Instead the entire school system seemed to be about learning technical knowledge, from math and physics in high school, to computer science in university. Then you go to interviews and you get technical tests, followed by some technical focus at the beginning of your career... but then after a few years once senior level expectations start to kick in... the expectation shifts and it's about learning how the business works, and how to make profit with software.

It was confusing why this shift happens in software. It might make sense in other industries where younger people need to replace older people (i.e. something with physical labor). So I'm now reevaluating how to look for a company that will leverage all of my existing technical knowledge or I will need to reconsider what to do because my career expectations don't align with the average software company's engineer expectations.

6. Ask HN: Writing an Interpreter in Go or Crafting Interpreters?

Top comment by stefanos82

Read both if you like.

I read and liked them both, but my heart belongs to Bob's writing, especially the C implementation in Crafting Interpreters and how he explains things.

7. Ask HN: I am at a loss. What shall I do?

Top comment by hardwaresofton

Put the startup on the back burner and go work a contract/full time job. It seems like you need the stability at least temporarily.

I disagree that you’ve fully failed (as the other commenter put it), but the current situation is untenable, you can’t make progress on your startup under that stress and you can’t survive/thrive in regular life.

Put that on top of grief from losing a family member and you’re basically doing the startup on hardest possible mode.

Unfortunately the job market isn’t great right now but if you put the energy you used to build the company in, you’ll find something.

Oh also talk to your investors and be honest/share your plan.

8. Ask HN: Where are people sharing their blogs these days?

Top comment by 1024kb

Some time ago @revskill posted a request for peoples blogs here: https://news.ycombinator.com/item?id=36575081

The responses were aggregated and posted to Github here: https://github.com/outcoldman/hackernews-personal-blogs?tab=...

You can download the OPML file and load it up into your RSS feed. I did start visiting the various blogs to find some i'd enjoy reading, but there are so many!

9. Ask HN: Cheapest way to host a back end

Top comment by toast0

You can easily get a VPS for around $5/month from quality vendors. You can probably find a lot of offers around $1/month for a VPS on lowendbox, lowendtalk, and webhostingtalk.

If you want more oomph than that, you can find dedicated server offers at those three places, starting around maybe $30/month. Typically pretty old systems, but I enjoy having a whole ancient box rather than one core in a vm on a modern box. My personal hosting needs are tiny though, so I could fit back into the VPS if money were an issue.

10. Ask HN: Did someone dig into the JFK files?

Top comment by didgetmaster

It is like when someone is being sued in court and claims in the press that they have turned over X thousands of pages of documents.

When you really dig into those documents you find 90%+ of them are completely irrelevant (old phone books, Chinese takeout menus, etc.) and that the few important ones that would actually shed light on the case, are missing.