4 ms·

The author notes that vibecoding has entirely replaced coding in hackathons (where speed is essential, bugs are tolerated, and only the demo is judged). I agree

The author notes that vibecoding has entirely replaced coding in hackathons (where speed is essential, bugs are tolerated, and only the demo is judged). I agree. But then says this means software is “solved” so only hardware hackathons matter. Why? If anything, I think software hackathons have become more useful, because ideas have become more useful. Even if ideas are cheap, not everyone has 24-72 hours for a prototype, in a creativity-inducing space that may inspire better details. And software isn’t solved: some ideas still require low-level knowledge and skill to translate into prototypes, especially if the hackathon judges require some functionality. Whether your purpose of a hackathon is: - Make a prototype, then if it seems useful afterwards rewrite it into a full product - Make a prototype that seems useful to attract investors (whether you start a company that may not launch or apply to a company that wants your creativity) - As an organizer, find ideas related to your company - Have fun, enjoy free food and good company

dvh 5mo ago
>But then says this means software is “solved” so only hardware hackathons matter. Why? Because ai cannot do hardware, it cannot solder wire, it cannot replace red led with blue one and find current limiting resistor for optimal brightness. It cannot see what part of enclosure needs to be cut. It cannot see the startup transient on the ldo.
- ssl-3 5mo ago
  > Because ai cannot do hardware It can. > it cannot solder wire Plenty of placement and soldering machines exist that are far faster than humans. They just aren't yet integrated with the bot. For the stuff that is unusual or particularly difficult, just add a human. Same as with code. > it cannot replace red led with blue one and find current limiting resistor for optimal brightness Sure it can. Just add a camera. > It cannot see what part of enclosure needs to be cut. Challenge accepted. > It cannot see the startup transient on the ldo. It can see that using the same tools we use to observe such things. https://github.com/aimoda/rigol-dho824-mcp https://github.com/aimoda/rigol-dho824-mcp
  - necovek 5mo ago
    And one needs to think from the outcome perspective too: maybe it is too hard to cut the part of the enclosure, but can it 3D print a new one hollowed out exactly where needed?
    - ssl-3 5mo ago
      Or, I mean: This is a [hypothetical] human hackathon that is tainted by bots, not a bot hackathon that is devoid of humans [although I'm sure we'll have those some day as well]. Instructions that deliberately lack specificity, like "Drill a hole for the switch and then mount it" work fine in meatspace whether those instructions are issued by a bot or a human teammate.
  - aiisjustanif 5mo ago
    Not at the entry price point of a hackathon, which is the point. Using AI in software hackathons are magnitudes cheaper than AI doing manual labor in hardware hackathons.
    - ssl-3 5mo ago
      > Using AI in software hackathons are magnitudes cheaper than AI doing manual labor in hardware hackathons. Sure: Going all-in on the bot with the hypothetical sorter and pick-and-place delta robot and the soldering machine? That's never going to be cheap. But one doesn't have to go all-in. The bot (potentially along with an army of subagents) can ingest a goal and a list of parts on-hand and their specifications. Like an army of bots that is instructed to write code, it can sit there and iterate on designs like a thousand monkeys with a thousand typewriters until it finds one that actually works. The humans (remember the humans?) can sit back and have a beer while this happens, and then start the physical process of hacking the thing together once the bot gives them direction. Look at this list of suggested concepts: - Building a ridiculous Apple II application - Turning a fax machine into a social media network - Turning a Game Boy Advance into a Bloomberg terminal - Making an LLM-driven cash register that can feel love and pain - An AI voice-activated microwave If a person can't imagine ways to use the bot for these things, then are they really trying?

dvh 5mo ago

>But then says this means software is “solved” so only hardware hackathons matter. Why? Because ai cannot do hardware, it cannot solder wire, it cannot replace red led with blue one and find current limiting resistor for optimal brightness. It cannot see what part of enclosure needs to be cut. It cannot see the startup transient on the ldo.

ssl-3 5mo ago
> Because ai cannot do hardware It can. > it cannot solder wire Plenty of placement and soldering machines exist that are far faster than humans. They just aren't yet integrated with the bot. For the stuff that is unusual or particularly difficult, just add a human. Same as with code. > it cannot replace red led with blue one and find current limiting resistor for optimal brightness Sure it can. Just add a camera. > It cannot see what part of enclosure needs to be cut. Challenge accepted. > It cannot see the startup transient on the ldo. It can see that using the same tools we use to observe such things. https://github.com/aimoda/rigol-dho824-mcp https://github.com/aimoda/rigol-dho824-mcp
- necovek 5mo ago
  And one needs to think from the outcome perspective too: maybe it is too hard to cut the part of the enclosure, but can it 3D print a new one hollowed out exactly where needed?
  - ssl-3 5mo ago
    Or, I mean: This is a [hypothetical] human hackathon that is tainted by bots, not a bot hackathon that is devoid of humans [although I'm sure we'll have those some day as well]. Instructions that deliberately lack specificity, like "Drill a hole for the switch and then mount it" work fine in meatspace whether those instructions are issued by a bot or a human teammate.
- aiisjustanif 5mo ago
  Not at the entry price point of a hackathon, which is the point. Using AI in software hackathons are magnitudes cheaper than AI doing manual labor in hardware hackathons.
  - ssl-3 5mo ago
    > Using AI in software hackathons are magnitudes cheaper than AI doing manual labor in hardware hackathons. Sure: Going all-in on the bot with the hypothetical sorter and pick-and-place delta robot and the soldering machine? That's never going to be cheap. But one doesn't have to go all-in. The bot (potentially along with an army of subagents) can ingest a goal and a list of parts on-hand and their specifications. Like an army of bots that is instructed to write code, it can sit there and iterate on designs like a thousand monkeys with a thousand typewriters until it finds one that actually works. The humans (remember the humans?) can sit back and have a beer while this happens, and then start the physical process of hacking the thing together once the bot gives them direction. Look at this list of suggested concepts: - Building a ridiculous Apple II application - Turning a fax machine into a social media network - Turning a Game Boy Advance into a Bloomberg terminal - Making an LLM-driven cash register that can feel love and pain - An AI voice-activated microwave If a person can't imagine ways to use the bot for these things, then are they really trying?