> What LLMs in Debian development will do, I fear, is eliminate any incentive to scrap boilerplate or reform policies that require a lot of other senseless human effort. If I had had access to LLMs 30 years ago, I might have just had them generate the rules files, replate with complexity. So they will make Debian even more firmly what it is, and ever less likely to explore what it could become.
I use and like LLMs, but it does feel like structural social changes are harder to do. For example, creation of new protocols. I hope I'm wrong, and LLMs are used to build collaborative, community projects - not just everyone vibe coding their own incompatible software.
That's perfectly wrong. Since strong coding AI, people venture into huge rewrites and other big changes that automatically make sense but otherwise would not.
Yeah it's a bizarre claim. Even better, I've literally taken some of the crappy experimental projects from the pre-Opus 4.6 era, and turned them into real software as models have become smarter.
I would say it's a mixed bag. On the one hand, yes, you can experiment with big changes much more easily, but on the other hand, you can have cruft build up in a big way if you're not actively looking for it and trying to cut it back.
In Debian, often the barrier to change is one or two people actively disagreeing with a change who are in a position to block it through their own action or inaction, so just being able to do more work on the change yourself is not necessarily going to make it easier to change things.
That's a very narrow view. If anyone could vibe code their printer drivers, maybe free software wouldn't have existed, because it would not have been needed in the first place.
Certainly, but if everyone can vibe code their printer drivers then we have free software in the sense of free speech and so users would be developers and the entire need is met. That is wonderful, not awful.
If it had happened early enough, code copyright might never have even been possible and no one can deny you your machine because you can just vibe up stuff.
Yes, I experiment much more. It's easier to justify, and, perhaps more importantly, just WAY psychologically easier to throw away a 20 minute experiment that your robot wrote for you than the same when it took you a week of tedious manual coding.
Yes, because LLMs have no online learning, it's a mistake to innovative in programming languages and tooling. Use what exists instead, the power of having training data is often underestimated.
His point is using LLMs to build software will lead to code so complex that it requires an Agent. Or that it will cause avoiding cleaning up the boilerplate because we can tolerate it now because of agents.
And thus this will all bias towards complexity and fragility..
Aside from Bun, which is owned by Anthropic, do we know of open source projects that allow AI where this has happened?
It's worth noting that Ted Nelson's visionary Xanadu project had actually addressed this issue decades ago. In the original design of Project Xanadu, rule 9 states "Every document can contain a royalty mechanism at any desired degree of granularity to ensure payment on any portion accessed, including virtual copies ('transclusions') of all or part of the document." This suggests a foresighted approach to content monetisation which the modern web, in many ways an unfinished hypertext prototype, still lacks. Nelson's idea was to embed financial transactions directly into the fabric of document interchange, an integration not yet realised in today's digital content landscape.
> It's worth noting that Ted Nelson's visionary Xanadu project had actually addressed this issue decades ago. In the original design of Project Xanadu, rule 9 states "Every document can contain a royalty mechanism at any desired degree of granularity to ensure payment on any portion accessed, including virtual copies ('transclusions') of all or part of the document."
It's easy to address every possible issue when your project is vaporware and you can just write "royalty mechanism" rather than having to actually implement one.
Yeah, wow. Saying that someone addressed the issue by writing "documents can contain a royalty mechanism" is like saying Jules Verne addressed the issue of manned spaceflight.
I've taken Amtrak. It's honestly not bad service wise. It is slow, has the lowest priority on a lot of track, has almost no backup options if something goes wrong on the train, and has some very old equipment in service.
If it were just twice as fast, was given priority or had it's own tracks, and had enough trains that a following one could pick you up in a disaster, and without random bedrooms that just smell like the 1970s, I would take it almost exclusively.
Otherwise.. sadly.. Amtrak is for when I don't actually care if I arrive at my destination or not. If I'm lucky I get where I'm going and most of the time I get a steak on the way.
You usually need to factor the whole trip as most trains start and stop in city centers while airports are sometimes very remote from the cities.
YMMV but where I live, I need to either pay for the parking at the airport, have someone shuttle me or take a taxi/regional train/bus to the airport.
For domestic use in my country, plane is competitive on paper, meaning some flights are much cheaper than train tickets. However the cheapest flights are the early and late flights which means regional train/bus to go to the airport are not an option. Usually you don't want to annoy friend/family very late in the morning/night because...well that's kind of rude. So you have to either use your own car and pay more for the parking than the actual plane ticket, or take a taxi that will also cost more than the plane ticket. Congratulations! The cost of your trip has already doubled or tripled from the advertised price! Now if you haven't thought of making your own sandwich add the aditionnal cost of buying food + drink at the airport and you will be adding again more than the original price of the one-way plane ticket! And that is without even mentionning the shady tactics of some cheap companies. Take a ryanair flight for instance, and once in a while they will ask you to check your luggage in one of their bagage check gauge and charge you whenever you have a bag that cannot stand upright on its own even if it fits the gauge and will force you to pay an aditionnal 70€ fee.
As for the time, it is also something where you have to consider the sum of all. My rule of thumb is regardless where I am flying, I know I will lose pretty much half a day. When doing sub 3h flights you are usually spending as much or more time commuting in/out of the airport and in the airport than flying.
On the other hand the train station will allow you to take your own drinks, and even if you didn't have time to make yourself a good sandwich, and don't want to pay for the more expensive options at railway station you will always find some cheap takeaways or supermarket close to it. The railyway station is usually in a city center as well which may be more convenient in many situations and you can reach the railway station 5 minutes from the departure time and you will still be fine.
Once you have factored all that, the train can often end up being the cheaper option, as well as the fastest one if this is an high speed train and you don't have to change train.
> as most trains start and stop in city centers while airports are sometimes very remote from the cities.
You write that as a negative, but for me that's just an even bigger bonus. The only thing I want from a city center is how to get out of the city center. I never actually want to go there.
I have not personally noticed much difference in commute costs plane vs train. Car trips, taxis, and public transport, have all been around the same time and cost to both. And the plane tickets I've looked at have been at all kinds of hours, early, middle of the day, and late. Exactly the same with the train, there's no specific pattern. (Obviously this is very specific to the location.)
You have a point with the food, it's easier to bring on a train - luggage as well, but then again the plane is so much faster you don't need to bring as much food.
And I do factor in wait time for an airport, very short trips are faster and cheaper in your car, longer trips are better in a plane. The train just doesn't really fit in anywhere - it's the same speed, or slower, than a car, and less convenient. (Even if the train is physically faster, getting to/from the station adds so much time the car works out faster.)
I book a lot of Spirit flights, and not only are they cheaper than a train, they are even cheaper than a bus! Every one in a while I go back and check all 3 options, and the plane and (occasionally) car win every time.
If you want people to take more train trips they need to cost 1/10 of what they currently do, or at least be 5 times faster. Without that, well, people just aren't going to use them.
More and more airlines are selling plane+train (or plane+bus) itineraries and giving the train system its own 2 letter value + “flight” number. Might happen.
