This is a great sign that uptake is low and someone needs a good Q4 on their performance figures for bonus season.
It was so (comically) bad that I ended up sharing a lot of screenshots on my process trying to make it work with my team just for entertainment value lol.
One time I asked it to do a diff between "col A" and "col B" and it started telling me it was a "goddess of light" blabla. I'm assuming because the random values in col A and col B completely polluted it's context window so it spewed nonsense back.
EDIT: I know the article is about the MSFT counterpart products. I've not used those with Copilot so can't speak to their quality.
But the goddess hallucination was the wildest and funniest one.
The dimensional merge is real! Did it also start talking about how it's a Hyperdimension Neptunia CPU?
I have never, ever, in my entire life, seen a tech work so hard to be everywhere while simultaneously not being very useful.
I feel like that one still hasn't really worked out, and I do occasionally use Bitcoin to buy things
In terms of web3, I think you could broadly say Bitcoin (though it was large before web3 so that one's muddy) and Etherium. But even then... given what they are, I'm not sure in what sense these are meant to be Goliaths...? It's alternative payment processors that have incredibly low adoption compared to virtually every other. They're "big" in the sense that they have a lot of traction relative to other crypto, but I still have never used the shit even once and I do not feel I am missing out even slightly.
I've used LLMs FAR more than any crypto, and I still see it largely as a neat way to get out of writing boring code, and a good rubber duck to bounce ideas off of when debugging. I wouldn't pay for it if I had to.
The only thing I know crypto for is being the new and preferred payment method for scammers, and that you used to be able to buy drugs with it but not anymore. And godawful avatar collections, I guess. I wouldn't call that a Goliath myself but. shrug
Are you not paying attention? Web3 is going just great![0]
It's remarkably fucking annoying and is genuinely one of the worst UX decisions Apple has made in like a decade.
No product in our entire history has been so aggressively pushed into everyone's face. If there's a person alive in modern society who hasn't had 4000 AI apps blasted at them... where are they and how do I achieve this nirvana?
Those things solved problems. Nobody needed to be convinced an iPhone was a good idea; it was an iPod, a PDA, and a newspaper stuffed into a device the size of a credit card that you carried around with you and worked no matter where you were. That's a GREAT idea. No one needed to be convinced about what made an iPhone (or Android, or even Blackberry) a useful and good thing.
Conversely, AWS made starting web businesses no longer require on premises servers, or really knowing anything about servers. You picked what you needed, and if you needed more at some point, you picked that. You could even dynamically allocate and deallocate servers on an incredibly widely available and robust data-center backend. That's HUGE. Numerous massive companies today may not have ever gotten started if not for AWS and it's now making far more money than Amazon's retail business does, and we know how huge it is because East-1 went down some years back, and a third of the freaking Internet stopped working correctly.
What problem even remotely on this level does an LLM solve?
It still took a couple of years, though.
Of course in hindsight its rather obvious. But there were other smartphones and they were (on paper at least) superior in quite a few ways.
I mean was that because of the iPhone as a concept or because it debuted as an exclusive on the worst cell carrier in the United States lol
You had Symbian, Palm, Windows CE devices which had all of that and keyboards. I won't pretend it was 100% obvious to me who was going to win at the time (and I don't think I was unique in that).
Translation problem. Google Translate does job poorly, if you compare with chatgpt. Especially if there are mix of languages in input.
Basic search of common, Wikipedia-level knowledge, to explain something. LLM is good at it.
- They have rapidly diminishing returns on higher contexts. They are very easy to overload with user information. You can post-train them, but that’s too much into actual computer science to be useful to the average office worker. They can work well as a co-worker assistant in a few cases, but they really can’t replace humans long term.
- LLMs only really work well when given the ability to call on-device tooling. Which, with a cloud system like Copilot, is going to be super tame and underwhelming because of lawsuits and various business deals. Tool calling is only really agnostic when you’re running the model yourself on your own device.
- On the topic of on-device models, there’s also the fact that AI provider companies have been caught evading things like robots.txt and causing so much crawling activity that it effectively becomes a DDoS attack. On-device AI doesn’t solve this totally, but most people likely won’t be pushing their gaming GPUs to the absolute limit 24/7 to constantly hit websites with crawl requests.
One 386 netbook runs xbuntu w/ firefox to waste hours in the dangerous wild except for HN. It's all a happy arrested-development `Smell of Teen Spirit' nook I continue growing in.
It should be a huge red flag if companies have to force stuff on users.
Reminds me of when I worked at a company once who hired a new marketing VP.
First quarter he showed up he hired a bunch of cronies.
Second quarter he announced that their new YouTube ads were so great now that NOBODY had ever chosen to skip their ads the entire second quarter. I saw the ads ... they had shortened all the youtube ads to the length that the ads were unskippable. Guaranteed "success".
Sorry if pointing that out is pedantic or rude, but it seems pretty material to the validity of the claim, at least to me.