When you have a (very very expensive) hammer, everything looks like a nail
On the other - the whole appeal of notepad was that it was a barebones text editor with none of that fluff (aka - it's a text editor, not a word processor or IDE).
MS has a large number of alternatives for the folks who wanted them.
When I opened Notepad - it's explicitly because I don't want the machine trying to tell me what I entered is wrong, or fix it. I just want a big dumb textbox for my file.
---
Basically - If you wanted those features you're looking for MS Word or Wordpad.
But MS discontinued Wordpad, and now they seem intent on trying to turn Notepad into Wordpad 2.0.
My prediction is this will not work well, since it directly competes with Office, and is not what the legacy users of notepad want.
But hey... a text editor is an easy place to shove text based AI, and cool shiny new thing of the year means some exec can claim to be shoving novel solutions into prod and bump up AI usage.
I’m very sceptical about all these AI announcements but text editing is case where I think this “AI” stuff can actually be used for good.
Anyway, I use TextEdit in plain text and autocomplete, autocorrect and spellcheck all work just fine, as they work in every text box in macOS. That Windows' Notepad got some of that just in 2024[1] is bonkers…
[1] https://www.theverge.com/2024/7/8/24194047/microsoft-notepad...
I assume this was the 1:1 correlated w/ the deprecation/removal of wordpad, as both were shipped everywhere prior.
[0]: https://support.microsoft.com/en-us/windows/windows-update-i...
why does Windows take 2 minutes to decide what updates to install, then 45 minutes to install them when Debian on the same machine can do both in under 30 seconds?
let me tell you a story...
It was supposedly worth all the power expenditure, because changing the world needed energy. Now we see where we are.
I'm inside this "newfangled AI thing". There are groups which create value, but they create value for everybody. The humans and the nature in general, and they use AI for scientific ends. Medical image processing, ecosystem monitoring, etc. etc.
Letting bots loose on the internet, letting them consume what they say and making them answer "Sauce is a food taste enhancer, and dressing is used to keep wounds clean while allowing them to heal. A standard serving of a dressing is two spoons".
Edit: although yes I do agree that the 'value' part is tricky. If internet spam can generate more 'value' for some people than doing science, then when intelligence is cheap we are in for a rough time.
Also, I'm very aware that there are many smaller models in production which can run real-time with negligible power and memory requirements (i.e. see human/animal detection models in mirrorless cameras, esp. Sony and Fuji).
However, to be honest I didn't see the same research on LLMs yet. Can you share if you have any, because I'd be glad to read them.
Lastly, I'm aware that AI is not something only covers object detection, NLP, etc. You can create very useful and light AI systems for many problems, but how LLMs pumped with that unstopping hype machine bothers me a lot.
I’ve yet to hear any good use cases for crypto, and I’ve been asking for years on here. Meanwhile there are a bunch of AI tools out there that are working and helping.
However, if we narrow what AI is to LLMs, we have a stochastic parrot which needs to be fed the world literally to enable it to create semi-coherent sentences about something being asked. More importantly, what that parrot says doesn't have to be true, it can't be guaranteed to be true, and can't be verified about its accuracy about its slop.
And you spend gigawatts of power just to train this thing which selects and prints words based on probability and some randomness.
That doesn't solve any problems.
Those "stochastic parrots" have still proven that they are immensely useful. You might not personally find value out of coding assistants, but many many people do (as an example). People are (allegedly) turning to LLM's rather than StackOverflow for help [0]. They work well for boilerplate where you're an SME and able to validate the output - I can review 10x the amount of code I can write for example. They work (remarkably) well for summarising input text. An example - I semi occasionally (3-4x per year) have to deal with a few hundred GB of audio files that need cleanup. The cleanup tasks are "run FFMPEG with parameters", except I can not ever remember the parameters (they're different for different things). I can: read https://ffmpeg.org/ffmpeg.html or I can ask ChatGPT to write a script to clip the silence and add a 0.5 second intro fade to every file in a specified folder, and the entire task is done before I've even thought about it. I get to focus on what I want to, rather than munging data around.
If you expand your definition from LLMs to Transformers, then you get Whisper as a stand out example of something awesome. There's definitely negatives, but things like Diffusion are being used outside of image generation for drug discovery. We're not going to yolo AI generated drugs into human testing, but we can save an awful lot of screwing around to find something viable.
> And you spend gigawatts of power just to train this thing which selects and prints words based on probability and some randomness. > That doesn't solve any problems.
I disagree, it does solve problems. A very fair question to ask is "is it worth the cost" and I would agree that it's not worth the cost. That doesn't mean it doesn't solve real problems.
[0] https://meta.stackexchange.com/questions/387278/has-stack-ex...
People are spending all that money training because they are trying to fix the problems you're complaining about, and this includes fixing the power consumption problem. If we can create 3B parameter models that have capabilities on par with today's 405B parameter models, that's worth spending a lot of energy training. But nobody knows what is possible, so they have to try. I feel like you're basically arguing nobody should try because you don't believe they will ever improve, but that seems contradicted by the general trajectory of how things have been working the past decade. More resources spent on training means more efficient and useful models.
Only if you ignore the fact that a database (in traditional sense) doesn't solve the problem of decentralized peer to peer payments, which is the key differentiator of cryptocurrencies.
> I’ve yet to hear any good use cases for crypto, and I’ve been asking for years on here.
If you've been asking for years, I'm sure that someone, at some point, has told you about crypto's censorship resistance and international payments in places that are poorly served by the banking system for a variety of reasons.
Would you like to hear more or have you already dismissed these as "not good use cases"? It would be nice to differentiate between use cases that don't apply to you personally, and use cases that don't apply to anyone.
And the point that the useful cases cede to make a useful product. The thing that is a _feature_ of a cryptocurrency is why people don't use it. I've had this debate dozens of times on here.
> If you've been asking for years, I'm sure that someone, at some point, has told you about crypto's censorship resistance and international payments in places that are poorly served by the banking system for a variety of reasons.
You know what else solves that? Cash and Western Union. And it has done for a long, long time.
People do, in fact, use them. Is it a popular payment method in western countries? No, but do some people use it? Yes, they do.
For privileged people, decentralization is usually a serious flaw. For others, it's an extremely important characteristic that lets them transact at all. The world isn't black and white, and people have use cases that are different from yours.
You're being self-centered, and that's okay, but perhaps you should factor it into your mental model before making sweeping statements in front of a global audience.
> You know what else solves that? Cash and Western Union. And it has done for a long, long time.
Not nearly as well, or there wouldn't be anyone using cryptocurrencies for that purpose.
You could make identical boring, bad-faith arguments about AI products. I think 99.99% of all "AI" products available today are completely useless - to me - but I don't go around proclaiming that all of AI is completely useless, and that all of its problem areas are better solved by statistics and "if" statements.
Don't mistake your own privilege, ignorance, and lack of imagination with the lack of real-world applications.
> You're being self-centered, and that's okay, but perhaps you should factor it into your mental model before making sweeping statements in front of a global audience.
> but I don't go around proclaiming that all of AI is completely useless, and that all of its problem areas are better solved by statistics and "if" statements.
> Don't mistake your own privilege, ignorance, and lack of imagination with the lack of real-world applications.
If you can't make your point without making sniping attacks about my character, then this isn't a conversation I want to continue having.
> For privileged people,
Privileged person is anyone living in a western country who hasn't had to deal with censorship. I consider myself to be a privileged person in that regard. That's not an attack on anyone's character.
> You're being self-centered
That's anyone who fails to consider use cases other than their own. I wasn't speaking to your character, It was a description of your reply, not your character, because it contained sweeping statements that only apply to certain groups of people.
> but I don't go around proclaiming that all of AI is completely useless, and that all of its problem areas are better solved by statistics and "if" statements.
That's not an attack on anyone?
> Don't mistake your own privilege, ignorance, and lack of imagination with the lack of real-world applications.
I've explained that privilege isn't an attack on anyone's character. As for the rest, sorry, but which words am I supposed to use when someone denies that a problem is real (which fair enough, I'll elaborate), later admits that there are other services that solve the same problem, but they still want to claim that there are no problems that the obscure product is solving, despite that product having real-world users who are using it for that exact problem?
AI is already implemented into businesses in various ways. Even if it’s not done so official you still have loads of employees pouring company secrets into chatGPT and Claude because they work.
It is certainly reasonable to suspect that the scale of investments (in trillions of dollars) don't match the scale of the opportunity. But it's a bit silly to pretend that no one is getting any value out of this.
At the end of the day, data centers are 2% of energy use, according to the IEA. That's trending up, but even in couple of years, data center stuff is mostly going to be typical cloud stuff, then crypto, and then a fraction for AI.
It would be order of magnitude more efficient to send couriers with cash to pay for those instead.
Never, in just 2-3 months we will have much, much bigger problems