AskHN Digest

1. Ask HN: I want to create IMDB for open source projects

Top comment by kpandit

I can only give you a datapoint but unfortunately no advice

1. I never discovered any movie on IMDB. I go to IMDB to find trivia, cast or some other fact about a movie that I somehow already knew of.

2. My interest in an open source project will not be influenced by its popularity or any other metrics but purely by what it means to me. I submitted my first PR to an open source project not because it is popular but because it lacked something I needed.

P.S. Thanks to all the nice people who generously contribute to OSS and offer their work for free. Hats off and respect.

2. Ask HN: What movies changed your perception of reality or life?

Top comment by bruce511

Groundhog Day.

It's such a subtle movie, that it works on multiple levels at the same time. I got to level 2 on the first viewing, but only years later assimilated level 3.

At level 1 it's a simple Romcom movie - boy stuck in a time loop needs to earn girl's love.

Level 2 is that we are all stuck in a time loop. Everyone is functionally living the same day over and over. Same job, same relationships, same everything. (There's even a scene in the movie where Phil describes his predicament and thd barfly replies "welcome to my life".)

Level 3 is even more meta. Because Phil is the only person -not- living the same day. All around him everyone else is the same, but he's evolving. We see him learning piano, and ice-sculpting and so on. He evolves emotionally, and is trying new things. He moves from being self-centred to focused on others.

The genius of the movie, and Murray's performance, is that it's buried, never forced. You're left to figure it out yourself. It's a work of art, disguised as something trivial.

3. Ask HN: Is RAG the Future of LLMs?

Top comment by gandalfgeek

#1 motivation for RAG: you want to use the LLM to provide answers about a specific domain. You want to not depend on the LLM's "world knowledge" (what was in its training data), either because your domain knowledge is in a private corpus, or because your domain's knowledge has shifted since the LLM was trained.

The latest connotation of RAG includes mixing in real-time data from tools or RPC calls. E.g. getting data specific to the user issuing the query (their orders, history etc) and adding that to the context.

So will very large context windows (1M tokens!) "kill RAG"?

- at the simple end of the app complexity spectrum: when you're spinning up a prototype or your "corpus" is not very large, yes-- you can skip the complexity of RAG and just dump everything into the window.

- but there are always more complex use-cases that will want to shape the answer by limiting what they put into the context window.

- cost-- filling up a significant fraction of a 1M window is expensive, both in terms of money and latency. So at scale, you'll want to filter out and RAG relevant info rather than indiscriminately dump everything into the window.

4. Ask HN: One-person companies—how do you manage it all and stay sane?

Top comment by mamcx

The major trick, IMHO, is not to grow.

While a "normal" startup is focused on growing! growing! growing! a small company should instead think like a physical restaurant:

"I have 10 tables * 4 seats. 4 waiters. Operate Lu-Fri 10 am - 6 pm. I can serve only up to 40 customers max per hour. My income will never be bigger than 40* Average meal. My income needs to be at minimum Costs + 3% profits".

Then you see you have too many waiters. You cut it to 2.

Then you see your meals are too cheap. You increase it a little.

Then you see you have little customers. You grow a little.

Then you see you have TOO MUCH customers. You stop growing. More costly take-over, increase the cost of the meal, etc.

What you NOT MUST DO is grow more than your max capacity.

That is it.

Size your capacity. Figure a nice profit. Keep it small.

5. Ask HN: High quality Python scripts or small libraries to learn from

Top comment by SushiHippie

I think I mention this all the time when this comes up, but I learned the most 'best practices' through using ruff.

https://docs.astral.sh/ruff/

I just installed and enabled all the rules by setting select = [ "ALL" ]

And then looked at the 'errors' it showed me for my existing code base and then excluded some rules which don't interest me.

I also have it set up in my IDE, so it'll show me the linting errors while coding.

In a larger code base there will definitely be many 'errors', but using it cleaned up my code really good and it stopped me from some footguns.

6. Ask HN: How do you ask users about their pain point?

Top comment by gnatman

In the realm of cold calling / emailing, 1% conversion is actually pretty normal. 2-5% would be unusually good. I would stick with what you're doing and don't lose faith!

If you want to try different strategies in parallel, I work in sales for a healthcare SaaS and while my product is results/outcomes oriented, I have much better luck talking to people about RESULTS they want to achieve vs. PROBLEMS they want to solve. At least now we're getting excited about something good vs. something they are already sick of talking about, or know they can't fix.

Some of the pains in healthcare are so structurally embedded in the industry, products purporting to relieve those pains are so plentiful, and ACTUAL pain relieving results are so few and far between, that messages looking for honest input on major pains that doctors experience is too clear of a dogwhistle for "I'm going to try to sell you a solution that doesn't work" and will be ignored.

Many doctors, for example, are still nursing wounds caused by the shift to EMR from paper- pretty much every vendor in that space promises an "easy migration" and the reality is that porting to EMR or switching vendors is a massive massive pain in the ass. ANY solution for a pain, or for amplifying/increasing a beneficial outcome, comes at not only a financial cost, but (perhaps more importantly) an opportunity cost associated with the time investment. Time = patients, patients = reimbursement, reimbursement = money.

Edit: Another thing that works for me is seeding the conversation with something like, "Something I hear from a lot of providers in [specialty] is they have a problem with [problem], or they want to do more [thing]. Is that true in your practice?"

7. Ask HN: Does it still matter to be in the Bay Area?

Top comment by someguy101010

I moved here ~2 years ago to join an early stage startup and while its "not what is used to be"[1] I have personally benefited a lot from being exposed to the business and engineering culture here. I'm starting my own company now and have access to much more networking events and people who "get it" than I think I could have access to if I lived anywhere else. I haven't though so its hard to say.

There is an incredible amount of tacit knowledge in this city that is hard to really internalize until you are actually here and living through it. I think programs like https://www.siltahouse.com/home take a great advantage of these things.

There's a lot of counter intuitive things about early stage that you can't really convey in a blog post, or you may not even believe until you see it yourself up close from people. The open minded attitudes and culture around risk is something that you can really internalize when you are around other people who give you the correct social proof.

Is this all things you can get other places? Sure! It may not matter as much as it did 10 years ago, but you will have a hard time not getting a positive experience out of living, working, or building in the Bay Area.

[1] people have said this since the beginning of time about everything anyways

8. Ask HN: What software sparks joy when using?

Top comment by brettkromkamp

Blender (https://www.blender.org/), by a wide margin, is my favourite piece of software. It's so apparent that the team behind it is passionate regarding what they are doing and highly skilled, as well. Each release brings both large and small improvements. What's more... it's open source.

9. Ask HN: Those who have recently undertaken a job hunt, what was your experience?

Top comment by spuzz

About a year ago, I sent out over a hundred resumes, got 3 callbacks, most of which weren't serious, and then there was one job I wrote a cover letter for because I matched everything they were looking for and handily got the job.

So it's both impossible and very very easy.

I think tech hiring is going to be a knife fight for a long time, because we're long past the point where only the thoroughly interested are looking for jobs in this industry, and there's so much money in the field that competition is high enough for people to take "towers of hanoi in optimized big o" interview tests seriously (probably because they cut the candidate field down to a manageable number)

>For those that have gone remote, any specific insights?

Be ready to manage your time, it's very easy to let 15 minutes off for chores turn into blowing off a whole day.

10. Ask HN: What older games are you playing through?

Top comment by galfarragem

Colonization[0] from 1994. Nowadays I play only the open source[1] remake: Freecol[2][3].

[0] https://en.wikipedia.org/wiki/Sid_Meier%27s_Colonization

[1] https://github.com/FreeCol/freecol

[2] https://www.freecol.org/about.html

[3] https://en.wikipedia.org/wiki/FreeCol