Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI is stifling new tech adoption?

vale.rocks

Post reply on HN
Print view View original

331–340 of 442 posts

Re: AI is stifling new tech adoption?

#331

Post by delanyoyoko »

Animats wrote: post #327
catapulted wrote: post #325

There is a counter example for this: MCP, a standard pushed by Anthropic, provides a long txt/MD optimized for Claude to be able to understand the protocol, which is very useful to bootstrap new plugins/servers that can be used as tools for LLMs. I found that fascinating and it works really well, and I was able to one-shot improve my CLInE extension (a coding agent similar to cursor.sh) to work with existing APIs/dat…

Now we're going to see sites specifically optimized to promote something to AIs. It's the new search engine optimization.

Prompt Engine Optimization

Re: AI is stifling new tech adoption?

#332

Post by ehutch79 »

graypegg wrote: post #292
janalsncm wrote: post #272

I’ve been out of web dev for a while, but maybe the problem is there’s a new framework every 6 months and instead of delivering value to the end user, developers are rewriting their app in whatever the new framework is.

This is obviously a pretty common belief, but I do think web dev gets held to a weird standard when framed as "the one with too much tech churn". Just because someone solves a problem with a new library or framework does not mean they solved a problem for all of web development, and I think the current concentration of applications made with boring things sort of reflects that. [0] > developers are rewriting their ap…

It's a lot better now, but there was absolutely a period when it felt like there was a new framework/library weekly. This was 5-10 years ago. 'Cambrian explosion' was the term going around.

A lot of it was in react space, with router libraries or the newest data store being the thing you needed to be using. Definitely turned me off react, personally at least. The angular.js/angular2 migration was also a big pain around this time as well.

There was a lot of social pressure from influencers on youtube and various social media that you NEEDED to switch now. This was the new hotness, it was blazing fast, anything else was obsolete tech debt. There was one instance of 'or you should be fired' that sticks with me.

I think we're just used to the hyperbole and are all a lot more jaded now.

Compare this to the backend, where django, rails, and the others haven't really changed. I haven't felt the need or pressure to rewrite my views/controllers/whatever at all.

Re: AI is stifling new tech adoption?

#333

Post by zkldi »

promiseofbeans wrote: post #296
datadrivenangel wrote: post #288

Earlier quoted context omitted.

Yeah but React is swapping major recommendations / styles ever 18 months still. How do you create a react application these days? Not create-react-app apparently?

NextJS for a SSR app, or Vite for a create-react-app style SPA. Vite has been great for the whole web-dev ecosystem, since it's super easy to configure, and very pluggable, so most frameworks are built around it now. That means you can write, e.g. a new CSS preprocessor, and if you make a Vite plugin, it works with (almost) every framework without extra effort

But this isn't even true, and NextJS is well into egregiously complexity. Remix was an alternative option in the space that is now deprecated in all-but-name for React Router v7, which (for those just tuning back in), react router is now a framework.

If you wrote your app in NextJS 2 years ago, you would already have to rewrite chunks of it to get it to compile today. These tools are NOT solidified, they are releasing breaking changes at least once a year.

Re: AI is stifling new tech adoption?

#334

Post by spenvo »

Kye wrote: post #323
jonas21 wrote: post #243

Earlier quoted context omitted.

You may be happy to learn that I just asked ChatGPT (the free tier that you can access without signing in) "What does the CurrentKey Stats app do?" and got this back: ---- CurrentKey Stats is a macOS application designed to enhance your productivity by offering detailed insights into your app usage and providing advanced management of your Mac's virtual desktops, known as "Spaces." Key Features: Time Tracking: Unlike…

The model and features available to it matters. o1 says "I’m not aware of a widely recognized or universally known product called “CurrentKey Stats,” so it may not be a mainstream or highly publicized app. There are a few possibilities, though:" and then some guesses. o3-mini is similar: "I couldn’t find any widely recognized information about an app specifically called CurrentKey Stats. It’s possible that:" When I t…

Fascinating. I would have thought the fancier models would have definitely known about it. The app is six years old, has ~120 ratings globally with ~4.5 stars, with several articles written about it etc.. It should be pretty widely available in the training data

Re: AI is stifling new tech adoption?

#335

Post by owenversteeg »

I think as new data gets vacuumed up faster, this will be less of an issue. About a year ago here on HN I complained about how LLMs were useless for Svelte as they did not have it in their training data, and that they should update on a regular basis with fresh data. At the time my comment was considered ridiculous. One year later, that’s where we are, of course; the average cutoff of “LLM usefulness” with a new subject has dropped from multiple years to months and I see no reason that the trend will not continue.

Re: AI is stifling new tech adoption?

#336

Post by ysofunny »

milesvp wrote: post #300
crazygringo wrote: post #252

No, AI isn't. Any new tech, or version upgrade, or whatever, takes time for people to become familiar with it. You might as well say "Stack Overflow is stifling new tech adoption" because brand-new stuff doesn't have many Q's and A's yet. But that would be a silly thing to say. I'm not going to adopt a brand-new database regardless of LLM training data cutoff, just because enough people haven't had enough experience…

Strong disagree here. I've been trying to learn Zig, and I'm thwarted enough by chatgpt giving me outdated information on Zig's unstable API that if I didn't have a strong incentive to learn it for it's cross compiler, I'd likely turn my efforts towards another language. This effect can greatly alter the adoption curve of a new tech, which can leave it dying on the vine. You're not wrong though, in that Stack Overflo…

or you could try leraning zig in the old-school way, the way we used to learn before LLMs

who am I kidding? LLMs have changed the game forever. making new computer programming languages is no longer professionally useful, like hand-knitting

Re: AI is stifling new tech adoption?

#337

Post by crabmusket »

ehutch79 wrote: post #332
graypegg wrote: post #292

Earlier quoted context omitted.

This is obviously a pretty common belief, but I do think web dev gets held to a weird standard when framed as "the one with too much tech churn". Just because someone solves a problem with a new library or framework does not mean they solved a problem for all of web development, and I think the current concentration of applications made with boring things sort of reflects that. [0] > developers are rewriting their ap…

It's a lot better now, but there was absolutely a period when it felt like there was a new framework/library weekly. This was 5-10 years ago. 'Cambrian explosion' was the term going around. A lot of it was in react space, with router libraries or the newest data store being the thing you needed to be using. Definitely turned me off react, personally at least. The angular.js/angular2 migration was also a big pain arou…

> There was a lot of social pressure from influencers on youtube and various social media that you NEEDED to switch now

I wish I had something more coherent to say about this, but: I think this is true, and it frustrates and saddens me that anybody took seriously what influencers on youtube had to say. It seems so obvious that they are there to get views, but even beside that, that they don't know anything about my app or the problems my users have.

Re: AI is stifling new tech adoption?

#338

Post by jimmaswell »

rhet0rica wrote: post #313
jimmaswell wrote: post #294

Earlier quoted context omitted.

QWERTY is a poor example. The keyboard layout is not the bottleneck for anyone who does not participate in olympic typing competitions. DVORAK is just as arbitrary as QWERTY to everyone else including professionals, and there's value in backwards compatibility e.g. old keyboards don't become e-waste.

It sounds like you're thinking purely about the speed of a skilled typist. Alternative keyboard layouts offer a tangible ergonomic benefit even at lower WPM counts, and can have a lower hunt-and-peck time for novices by clustering frequently-used letters together. (This last effect is particularly pronounced on small touch screens, where the seek time is non-trivial and the buttons are much too close together for any…

I remember reaching a really high speed on the keyboard of the original iPod Touch. It actually did feel like touch typing - I didn't really have to look at the on-screen keyboard. I can't pin down exactly what's been missing from newer keyboard apps. Something about the ergonomics and UI came together just right.

Re: AI is stifling new tech adoption?

#339

Post by feoren »

The answer to this seems obvious: continuous training of live models. No more "cutoff dates": have a process to continually ingest new information and update weights in existing models, to push out a new version every week.

Note that I said "obvious", not "easy", because it certainly isn't. In fact it's basically an unsolved problem, and probably a fiendishly difficult one. It may involve more consensus-based approaches like mixture of experts where you cycle out older experts, things like that -- there are dozens of large problems to tackle with it. But if you want to solve this, that's where you should be looking.

Re: AI is stifling new tech adoption?

#340

Post by thatguysaguy »

feoren wrote: post #339

The answer to this seems obvious: continuous training of live models. No more "cutoff dates": have a process to continually ingest new information and update weights in existing models, to push out a new version every week. Note that I said "obvious", not "easy", because it certainly isn't. In fact it's basically an unsolved problem, and probably a fiendishly difficult one. It may involve more consensus-based approac…

Yeah I think every lab would love to do this and the field has been thinking about it forever. (Search lifelong learning or continual learning on Google Scholar). I don't think a technological solution is likely enough that we should pursue it instead of social solutions.
Post reply on HN