Comment by levkk
1 year ago
If my Android (or IPhone) disappeared tomorrow, I would feel like I time traveled back a century. If Google search was gone, I wouldn't be able to do my job anymore. If the cloud disappeared, I wouldn't be able to build apps anymore. There are no workarounds, unless you feel like going to a library...?
If ChatGPT disappeared tomorrow (or derivatives like Copilot, etc.), I would be mildly inconvenienced. Then I'd go back to reading docs, writing code slightly slower and carry on. In fact, I did this already, several times (Copilot with GPT-3.5, Cursor, Copilot with GPT-4, Zed with Claude, etc.)
I think that's an unfair comparison. If the IBM Simon disappeared in 1994, I'm pretty sure you wouldn't have cared. If search engines disappeared in 1992, you'd have felt the same. Also, (what later became) AWS probably didn't interface much with you in 2003.
It takes some time for technology to mature, usually at least a decade or two. Even once the iPhone was released it took a few years until it became indispensable.
If IBM had been spending hundreds of billions on its phone thing in 1994, though, well, there would probably have been no IBM by the early-noughties when smartphones that people wanted to use started to become practical.
"It may or may not produce something useful, in a few decades" cannot justify the present level of spending; that's just not going to fly. Without concrete results _soon_, the whole thing is in very big trouble.
Also, it takes time for people to forget what it was before so when their current status quo is taken away, they don’t know what to replace it with.
I remember when people commented here that the blockchain was the same as early google search or early aws or early iPhone.
Everyone thinks their new thing is the T-1.
That's always been true. We went through PDAs, 3D TVs, smart cities, cyberspace, but some of those actually did become our iPhones. "Nothing ever happens" is just hindsight/survivorship bias.
1 reply →
I spent a moment with my brain froze up apparently trying to remember when the first Terminator was supposedly deployed in one of the timelines.
Then I realized the reference was to the "Model T" car. Somehow in my brain the token for "Model" is actually necessary for the correct lookup.
And yet many, many people - including me - were shouting from the rafters that this is nonsense, that Blockchain has no real use case (aside from cryptocurrencies), and that this was all a massive bubble.
Most of those some people - again, including me - are saying the opposite about AI.
The lesson to draw isn't "there's a lot of hype, means there's nothing there", nor is it "there's a lot of hype, means there something there". It's "we need to actually think about the technology we're discussing and make object-level decisions, not meta-level decisions".
But the AI-public Internet timeline is more like 1995. if all web search disappeared in 1995 it would've been a massive loss, despite how primitive search engines were back then.
reminding me that ~1995 for a couple of years it was practical to print listings of web sites:
https://www.goodreads.com/book/show/2868341-the-internet-yel...
Chat right now is search in 99. Highly functional
Did you read the article? The author spends a LONG time going over why it's not early days for LLMs.
Here: https://www.wheresyoured.at/longcon/#:~:text=Also%2C%20uhm%2...
If Google search disappeared, I would still remember the names of several sites like Stack Overflow that are always at the top of the results, and I’d just go directly there.
Maybe the original Yahoo! style curated list of categorized links would actually be more useful for me at this point than Google with all the SEO spam.
That kind of high-quality directory combined with ChatGPT would probably replace Google for me.
> If Google search disappeared, I would still remember the names of several sites like Stack Overflow that are always at the top of the results, and I’d just go directly there.
I kind of agree with this, but Google is still, IMO, the best way to search these sites even when you know they exist because most of them have terrible local search.
Google searches with the site: tag are one of the few ways in which I find google search to be somewhat useful. Its pretty terrible these days at more generalized search due to their capitulation to SEO.
I think this was true for a long time and so it's burned into a lot of us as conventional wisdom, but have you actually tried lately?
I've made a habit to go directly to the website I want and use its search first. In almost all cases I find what I want, and I get to avoid touching Google.
Does anyone maintain a Yahoo-like directory anymore? I used to love browsing the internet that way, and I feel like it would be a really interesting way to find websites rather than content.
I agree, but I think such a project is just not realistic anymore. If you make it publicly editable (like a wiki) it'll just be full of spam, and if you don't, then you need swathes of human editors. And then you get into endless fruitless debates on exactly which websites should be listed and which should be excluded.
I've had some luck googling "awesome" "<topic>" "github" -- people make lists of projects and links and papers and such and I've found some gems on there when I'm looking for dataviz or csv cli tools or what have you.
Before 2015, I didn't have a reliable internet access (as in I can access when I want to, not I have access all the time) or electricity. But I manage to learn much about computers and programming. Why? Because a book is actually a very dense information repository. And the rest, you discover by reading code and manuals. The reason I use the internet today is mostly because no one takes time to polish their library and document it. Instead you have to wade through issues report and forums. Compare that with C, where it's much more slower and you can use man for function names.
The internet is much more faster, but you can do stuff locally if someone took care to download documentations (which I did because of the above reasons).
To play devil's advocate here, it takes time for society to orient around new technologies and make them feel indispensable.
If smartphones had disappeared in 2008, most users would be mildly inconvenienced. They'd go back to using a flip phone and a TomTom, or printing out map directions, and sending emails on their laptop with WiFi. No employer expected them to have a chat app on their phone* or use PagerDuty, no businesses required them to download an app to purchase services. People called taxis on the phone.
Perhaps in 5-10 years people will stop putting any effort into documentation or organizing information, (some companies are already ahead of the curve on this one) and our jobs will become that much harder without an LLM to sift through all the information.
* I'm ignoring the BlackBerry world here which was always pretty niche.
> Perhaps in 5-10 years people will stop putting any effort into documentation or organizing information
The LLM needs the documentation more than I do
LLMs (or other AIs) of the future may well get better than humans at reading badly written documentation, filling in gaps by cross-referencing multiple sources, or learning APIs entirely from reading uncommented source code...
I agree (mostly), but
> If the cloud disappeared, I wouldn't be able to build apps anymore.
Seems rather hyperbolic. Colos and shared hosting were a thing long before the cloud craze, and still continue to be a thing. I figure if nothing else a lot of people would go back to the still relatively low touch environment of uploading PHP or CGI scripts, which honestly seems like a pleasant change these days.
I don't think Google search is useful, as a developer. The top results are all garbage, unrelated, or ads. Going past the first page used to be useful, now it's not. Honestly, it's so bad, that I straight up do not believe you that you use Google to find new information.
DuckDuckGo/Kagi are where good search is at.
Agreed, I don't even bother with Google anymore. The only thing I use Google for is for local results, e.g. finding a restaurant. It's useless for finding actual information on the web.
>If ChatGPT disappeared tomorrow (or derivatives like Copilot, etc.), I would be mildly inconvenienced. Then I'd go back to reading docs, writing code slightly slower and carry on.
Except this is exactly what you would do if Google disappeared and did before Google existed. You're applying different standards.
While that might be true for you, it's not at all true for me. If generic chatbots went away, no big deal. If Windsurf IDE + Sonnet 3.5 went away tomorrow, my prospects would look very different.
The last production code I wrote was over 20 years ago. I don't know React and TypeScript. I recently created a SaaS MVP using Windsurf/Sonnet/React/Refine.dev/Supabase in 8 days. We already have live humans excitedly using the product.
The SaaS product is a recreation of an app that I tried as a startup a few years ago, which never got to even this level of traction and failed. We failed for many classic reasons, but one of the main reasons was that we had no truly technical co-founder, and could only afford an off-shore dev. Iteration took around 24 hours. Using Windsurf, a product shmoe like myself can iterate in 2 minutes.
Of course we will have to get a real React dev on board if we start to get real traction, but the LLM tools allowed us to explore an opportunity that would not have existed without them.
Disclaimer: I happened to use Windsurf, but there are other options like Cursor, which you might have better luck with. I am not on Mac OS, so Cursor was not an option for me.
What remains to be seen if your venture (and others like it) will translate into hundreds of billions of new economy. That was the effect of the iPhone, cloud, and Google Search (analogies used in the article). That's the difference between a "new era" and a cool tool that could of (and should of) been built at university with enough funding to produce it for the public good.
Yeah, I can agree with all of that comment.
My gut feeling is that this will play out like the dot-com bubble. There was definitely a bubble, it burst and many investors lost money, but the Internet did end up eating the world in the long run.
Congrats on the rapid MVP! Is it launched / publicly available? I'm in the same boat with modern front ends. Would like to see what the current LLMs can help with.
Thanks! It is not public yet. It's a b2b web app + browser extension. We currently have users in a friendly company testing it to replace their existing cumbersome Excel + Email/Teams workflow.
If you are in the same boat as me with modern front ends, come on in, the water's fine! I highly recommend using something like Refine.dev (YC S23) + Ant Design, or whatever is appropriate in your case. Giving a tool like Windsurf/Sonnet a much more narrow scope of options than just "React" is a huge win.
Even if you just started from a paid or free template for whatever language and framework you are targeting, it will greatly improve your chances of making something appealing, very quickly, using a tool like Windsurf.
Google search example is interesting I wouldn’t care at all. YouTube or maps on the other hand…
I've actually been wondering what platform(s) people would flock to if YouTube were to suddenly disappear. Vimeo? DailyMotion? Or even... PeerTube??...
It is sad that searching online is so much more convenient, since between downloading dumps of Stack Exchange sites and downloading the full documentation for programming languages and libraries I use it would be easy to have 99% of all the answers saved locally, but it would still be faster and easier to look those same answers up in the cloud.
And generative AI made it even worse, probably worse than any single thing before it.
Now you're being fed incorrect answer by the search engine built in llm, it's impossible to find legit reviews: they're either sponsored reviews or written by bots, image search is next to useless, it's impossible to find a recipe, you can't tell if they're legit or if they're written by an LLM and will completely fall apart because as it turns out the best cookie recipe isn't the average of all known cookie recipes
People lose their phones all the time. No one has a dramatic time travel moment. You open up your laptop, or walk to your desktop. It's not a necessity excepting actual phone calls.
And how much money per month are you willing to pay for each of these capabilities/services? I can't imagine paying $50/month for any use of AI.
I willingly pay $80/month for a smart phone and its ecosystem. Google search (for general queries) is worth a lot less to me, since it's fairly easily replicated by federating a search across a dozen sites where most answers arise now. So I might pay $5/mo for that, or maybe $20/mo for code queries (to be paid by my employer of course). Internet access for desktop/laptop computing or for media streaming might be worth $40/mo.
So what are the end uses of AI that I'm willing to pay generaously for? If GenAI is supposed to revolutionize the infosphere, the question has to be, what service will it provide that can justify its cost -- the $trillion investment in infrastructure that is underway?
I have absolutely no idea what AI's killer app could be. IMO, not even a dedicated secretary/tutor/companion is worth more than $50/mo to the average bloke. And that drip of revenue per user isn't nearly enough to justify the development costs that AI is incurring now.
If Google search disappeared and ChatGPT was still around, you could replace Google search with ChatGPT and still do your job
That's why it's not a completely product hype driven bubble since there is a useful product there
I think the above poster was comparing search engines to LLMs and just using Google as a generic example.
LLMs very obviously aren't a replacement for live human written information which is findable with a search engine.
For me Google search (and YouTube) feels like its downgraded in recent years, even before the AI craze. Maybe people aren't using SO as much as they used to.
If electricity disappeared we would go back to lighting candles…
And clearly candles are definitely viable products.
[dead]