Hacker News Reader

Comment by axegon_ | Hacker News Reader

Makes perfect sense, all things considered. I've only joined a handful of hackathons. My best experience was in Amsterdam in like 2022, where half our team went to sleep and me and another guy spent the entire night locked up in a venue with 200 other people building stuff and bashing our heads against the table, looking for optimizations, hacks, half-assed solutions to near-impossible problems. In recent years, I've lost interest: And at this point I don't think I'll ever join another one: I recently got an email about one that finished and the winner was a guy who created something like an "AI team of engineers". What he presented was 20 markdown skills.md bs files. I mean seriously, being literate is enough to get you the gold medal? As a friend of mine likes to say, "you hit rock bottom and started drilling into the rock now".

At least with hardware, people are actually making something and have to use their brains.

View on HN

Most hackathons have been this way for a long time. I recall spending a weekend working out some really thorny data classification problem but got nowhere, all the winners had slick presentation slides and a 30 second code demo of a glorified CRUD app.

> A bad design with a good presentation is doomed eventually. A good design with a bad presentation is doomed immediately. - Akin's 20th law of spacecraft design

I always really enjoyed making a slick presentation. It was a lot of fun figuring out how to scope the hardest problem you are sure you can finish in 24hr while still having time to polish your presentation and make the app look good. I find picking a problem that lets you put a big map on the screen helps with the latter.

I get that… but that’s basically a startup pitching competition. It’s not a hackathon.

Aren't most hackathons pseudo-startup pitching competitions? At the very least, they've always been about established companies trying to extract value from newcomers.

To a large degree, this is how the real world works too.

Absolutely. Which is why the hackathon was conceived: to serve as an escape from the real world.

All good things must come to an end, I suppose.

When I first read about the hackathon concept, it was a bunch of OpenBSD developers getting together to work on stuff like cryptography or drivers. I think it was from Facebook where I first heard it as some bullshit corporate event. The idea of a "winner" seems to misunderstand what should be a goal. It's not a competition. It's collaboration.

I’ve always heard of hackathons as being a competition with judges and a winner.

Sounds like you only know the corporate or post-commercialized version, and not the original concept. The entire point of my comment.

And it drives the mini-MENSAs of the world insane.

10 years ago I won the only hackathon I participated in (not by choice). The jury was especially impressed in our report with the AI section. That part was a bunch of technobabble that I wrote, more or less saying "in the future the system should do this and that", quoting some popular algos from that time. None of it was implemented in our demo in any way, shape or form. They checked that I knew what I was talking about, and talking with confidence was all it took.

We did not even try to win the hackathon, just to get a passing grade.

I mean, this is hardly surprising. Who takes the most points is an accumulative score from subjective opinions(judges, audience, etc.). We didn't win the one in Amsterdam but got second: Around 50 teams began, 20 managed to deliver something, even if the winner is picked at random, that's a 5% chance, which is a very high random chance. When you toss in several senior developers(who at the time worked together in the same company and team), a dedicated frontend developer, ux designer and a few others, second place no longer sounds that impressive, but we all had fun. But to my mind, the value of hackathons is(or rather 'was', given what I said above) forcing people to push their mental abilities to the limits. If being able to write coherent text is good enough to make you the top performer, then we clearly have a problem.

My point is that flashy presentations already were an issue before the rise of LLMs. The evaluation of a hackathon relies on presentations and subjective opinions rather than pure benchmark of technical assets, there is nothing new under the sun here.

You can now win with 20 skills.md files now, you could win with "it would be great to use this sexy tech" 10 years ago.

Unfortunately that will be short lived too. There's a lot of people toying with LLMs to develop hardware without understanding it too :|

A lot of my personal time, and a few others that I know, is being spent either updating, replicating or simulating the classic BBS experience today... while not really simulating the Modem itself.. trying to get the way a certain terminal looked in the old days is a bit of arcane knowledge combined with the time to make it work.

AI does a pretty good job of a lot of it... I've mostly been using Rust for my target language, the biggest parts so far in terms of rework/retries has been getting the "experience" how I want it, which isn't really a 1:1 of the classic experience, but updated. Like mouse scrolling for scrollback controls, etc.

There's also been a lot of activities in enhancing the BBS software that remains today (Synchronet, Enigma, Mystic and a MajorBBS remake) along with continued classics and some door games.

True. And it has the same safety implications as in software and security but more obvious since the slop creeps it's way into the physical world: One of my main hobbies is drones. Lately less flying cause of a long track record of crap weather every weekend for like 6 months so I'm more on the building side of things. I would not trust a slop machine to design and build something that weights 4 kilos, carrying a highly flammable lithium(ion or polymer) battery and fly it in a remote field even. Off the top of my head, I can think of 200 ways this could go wrong. And most people with a functioning brain that have watched the news will agree with me. The line between "Claude sloppus is so good man, look at this awesome thing it did" and "Lost control of the drone, it flew into an airfield and crashed into the engine of a plane taking off" is incredibly thin. What pisses me off is that if this happens, regulations will hit those of us that know what they are doing and will make sure this never happens, and not the geniuses who think slop is a viable solution for everything. Same story with the never ending leaks and supply chain attacks which are a direct consequence of sloppers.

> What he presented was 20 markdown skills.md bs files. I mean seriously, being literate is enough to get you the gold medal?

Yes? If your problem is that there a tree in road and one guy builds a autonomous robot to remove it and the other guy just goes and moves it, the “dumb” guy wins. We are at a point in history where a couple of markdown files solve problems better than hundreds of hours spent by experts in building dedicated solutions. But you win based on the results not based on how much effort you put into it.

It's what I call "speed running towards watering plants with Gatorade".

You suggest doing it with water, perhaps? The same one we use in the toilets? Yuck!

But the winner didn't remove the tree. They just built a an AI team of tree-cutters...

That may not actually cut trees.

Yes. They wrote an AI prompt to cut a tree and gave a presentation about how this is the future of tree cutting. They did not cut the tree, nor build a robot that cut the tree. They fantasised about building a robot that builds a robot that cuts the tree.

Hackathons are about making and finishing something. That’s the whole premise: Instead of spending months working on a personal project that goes nowhere, give yourself a time constraint so you’re forced to finish something. Now you have something concrete to put on your portfolio (even if it’s still a proof-of-concept) - plus you’ve probably learned a lot.

Setting up config files for a tool is not making something. That’s like if I spent hours setting up my IDE, build processes, a CI/CD pipeline, and even unit tests. That might be cool and enhance my productivity down the line. But I still haven’t made anything.

> Setting up config files for a tool is not making something.

This is like complaining if someone used Unity at a game making hackathon. “They didn’t really bill anything they just configured an existing solution!”. At the core hacking hackathons are about what you deliver not the effort put into it. It’s crazy how people want to turn it into a purity question. People feel threatened because AI has removed a barrier for those whiners not as good at programming as they see themselves. But hackathons where never won by the rules of people who go “I am going to build my own game engine because using existing tools and libraries feels like cheating!”

What is the gold that were all mining for?

Ah yes higher fidelity, external stimulus + human response tuples to feed the ad revenue machine

Reply on news.ycombinator.com