Yeah that’s a close enough comparison.
They’re both greedy, to say the least, but at least when Nintendo releases a new console it is actually different (at least on the power side) from the predecessor.
1 Like
By the way everyone, thanks for the responses that you’ve been giving! This topic is so interesting to read and there is a lot to take in, so sorry if I take a while to respond to everyone.
2 Likes
That’s because every Nintendo console exchanges power for a gimmick (with some being MUCH better and useful than others).
I’m not sure what gimmick they’ll pull after the Switch 2 though. I mean, the Switch 2 is literally just a more powerful Switch.
1 Like
Well yeah, the switch 2 is a better switch 1, but the gimmick remained basically the same except the Mouse-Joycons.
The best thing the could do is bring back the Dual Screen format while keeping the detachable joycons thing, then they’d have a console which (if worked on properly) could be made basically backwards compatible with the whole Nintendo library through adapters of sorts.
But I doubt they’d do something similar, they definitely prefer having retro games on their Nintendo Switch Online subscription.
I heard they did a patent for an attachable screen or something though, that might be me misremembering.
1 Like
That’d be super cool, but Nintendo seems to be allergic to game preservation and accessibility.
What other games get MORE expensive the longer they’re out!
1 Like
Yes, exactly.
1 Like
I think that you and others who didn’t go through it just cannot comprehend the similarities because the internet has been such a huge part of your life. The DotCom bubble was absolutely driven by FOMO.
In March of 2000, the DotCom bubble burst. For the previous few years, every company had been investing in web pages. You couldn’t be left behind. Everyone had to have a web page. Business were paranoid about not being ready for the future.
Shopping, Entertainment and Social Media
Online stores existed, but Craigslist and E-Bay were the most popular place to buy things. Amazon was known as an online book seller. It was trying to expand into other markets with limited success. Netflix was a mail-order DVD rental website - it would be another 7 years before they started streaming videos online. Youtube wouldn’t exist for another 5 years. Facebook was another 4 years away, and even then it was only available to students - it was another 2 years before the general public even heard about it.
Searching
Google had been around for 2 years, and had just started selling ad space - it was starting to edge out other web search engines like Lycos and Yahoo. Most of the world used Windows computers with Internet Explorer as an embedded browser - the anti-trust lawsuit that would prevent them from having a monopoly on browsers was still a year away from concluding (it started in '98.). Chrome wouldn’t exist as a browser for another 8 years.
Hardware
The iPhone was 7 years away from even launching - at which point it would be a luxury product and would make the general public aware of smartphones. NVidia had just released the GeForce 256 - the first video card with hardware acceleration the year before. It put longtime leader in the the marketplace, 3dfx out of business (NVidia acquired them in 2000).
Internet Providers
Cable internet had been out about 4 years. DSL internet had been out for about two. But those infrastructures were still being built. For those who had it, it was a luxury. So most of the world who even had internet was on dialup. Which meant going to a webpage was comparable to waiting for a 250GB game to download today. Wireless internet of any speed at all didn’t come around until 2009 with the iPhone3Gs - the infrastructure was still being built.
Most people who were on the internet used their phone lines, and went through AOL. AOL and Compuserve (which AOL bought that year) had such a huge intranets that many users didn’t even venture out onto the actual internet. You could shop, talk, search, etc there.
Average User
Think about your family members who struggle with using their phones and computers now. None of them were online in 2000. Most of the world was not online. Commerce was not driven online - but everyone could see that it would be. Students were not online unless they were in college and even then - it was mainly computer science and engineering students. Video games were for geeks and were still primarily targeted at boys and young men.
The average person, if they were online, treated it like a toy - not a way of life. Elderly people weren’t online at all for the most part. According to the U.S. Census Bureau in 2000, 51% of households in the U.S. had a computer in the house and only 42% had internet access. Compare that with today, where last year 96% of people in the US are connected to the internet.
Stock Market & Money Today
Apple was started by two guys in a garage. Microsoft was started by a guy in his mom’s garage. Amazon was started by a guy in his garage. Google was started by a couple guys in college. Facebook was started by a couple guys in college. OpenAI was started by a bunch of rich people with a billion dollars starting capital and is the outlier. All these companies though, saw how the lack of infrastructure caused the DotCom bubble burst - and so they are all investing in infrastructure to try and prevent that from happening. That is why NVidia was the first company to be valued at $1 Trillion on the stock market - everyone is buying from them.
We’ve discussed in other threads how the current ROI for use of LLMs is 3% of companies. And of the big companies I mentioned two paragraphs previously who are throwing billions of dollars at this - only one has made any money. And that of course, is NVidia. These are HUGE loss leaders. And while the big companies can suck those up -most companies cannot.
Right now the DOW, NASDAQ and S&P 500 are all heavily skewed by those big multi-trillion-dollar (valuated) tech companies. Which means that while we not have a stock market crash, companies smaller than a trillion dollar valuation can crash en masse and not affect those numbers anymore. So while the “bubble bursting” may not look exactly the same this time around - it could still have profound economic impacts.
LLMs Today
Today about 47% of people in the US have used LLMs, with 29% using them weekly. (Full disclosure: I am one of the weekly user myself.) Those numbers are markedly similar to the usage numbers at the time the DotCom bubble burst. People know that LLMs are going to be big. But like the late 90s with the internet, people don’t know how. So investors are throwing money at people - just like they did in the late 90s.
However, what they cannot control is adoption and public perception. And they are hitting HUGE roadblocks that didn’t exist in 2000. Namely, the environmental impacts, the ethical implications, and people visibly losing their jobs en masse. The money being poured into the infrastructure is also an issue.
Conclusion
For those of us who were around (and working) in 2000 - the rise of LLMs have a LOT of parallels with the rise of the internet and world-wide-web Keep in mind the internet had only been available to the general public at large for 5 years. In comparison, LLMs have been available at large since the launch of ChatGPT on Nov. 30th, 2022 - 3 years.
While adoption numbers for LLMs are slightly higher, and the technical hurdles for people accessing LLMs are much lower - the ethical and environmental impacts are limiting general adoption by individuals and some companies and industries. The implementation of LLMs is much more controversial than internet adoption - because there was very little resistance to the internet in comparison.
Regardless, the underlying issues are the same. Companies were and are investing in a new technology that they could not afford not to - but nobody knows how to monetize it yet.
6 Likes
Seriously ‘AI systems integration’ could be a job, perhaps a conpany that turns up and plugs it in.
Then a training course would help the member of staff get used to using the AI, which is of course transfer learned into its position …
And then the successful company can almost be automated to self-run. Theres no hassle, the company consumes AI tech, an internet connection, training for the member of staff, and installation services.
Why waste any more time coding?
Wow. I didn’t expect all of that.
I kinda knew that many companies were investing in websites in fear of being left behind but I had not considered the status of the internet infrastructure at the time that made the Investments really risky.
It really does look like the companies are trying to avoid another burst, the infrastructure is definitely in different conditions than the one the Internet had back in the day but the tech is still very precarious.
There are definitely many similarities, but many differences too.
Apple was started by two guys in a garage. Microsoft was started by a guy in his mom’s garage. Amazon was started by a guy in his garage.
Are you suggesting that we should all go in a garage and something miraculous will happen?
I loved your response, thanks!
2 Likes
I’m sorry to burst your bubble (pun intended), but having an AI that is competent enough to fully replace your coding efforts is currently a pipe dream.
Even if an AI could fully code for you, it still wouldn’t be able to read your mind and do things your way. Unless we created a digital clone of you, but nobody’s begging for that.
1 Like
Well, that is vaguely what is happening. LLMs are being used as a foundation for new services in many fields. It’s not one company that needs an integrated system, but a whole set of companies virtually doing the same thing - and they all are targeted by new services that they ought to be using.
And of course there are offers of plain old “we do it on your site” services, but those individual solutions are usually more costly for the targeted customer (company) and require a long time to be implemented, as it is usually about building it on top of the already available infrastructure (think data and systems) instead of a new “just use this service from now on”.
The tricky part is the balance of losing established jobs vs. creating new jobs. No taxes means no social balance. No income means no taxes and no consumption. No consumption means no revenue.
3 Likes
I think a human could achieve concious immortality in digital form before any software as a service people could concede the possility that the computer was better at scripting. That idea is for the far off future.
For the present day I would be considering whether the version of the AI that i could use was a dumbed down version of the actual in house one the big companies can use.
If i ask the computer for a blender python script then in less than a second it solves the problem and expresses compicated linguistic structures that i would take some time to create. I would share one with you, but the forums dont allow it.
One thing I notice a lot in talks about this topic is this argument. Just upfront: I don’t mean to degrade this in any way - it is fair enough point to make.
LLMs work quite well for generating solutions for single-problems, sometimes even step by step to a bit more complex situation. But usually these are problems in the lower complexity sector, isolated and disconnected from a bigger context.
The problems for me begin with the illusion of LLMs being able to do more than that. Maybe it is too soon to ask for it but does anyone actually know a bigger project run by a small team or even individual sporting a LLM to do the coding? Something that does integrate into other systems, too? Because I have so far only heard the opposite (which was brought up a lot already in this thread) about projects becoming unmaintainable.
(Sidenote: is there a new term for spaghetticode generated by a LLM? Or is it just slop?)
3 Likes
I dont usually use an LLM for scripting, but i always use a web search if i need to know about something. The step by step method was always the best way to do anything that hadnt been done before. Unless you invent your own Djikstra or puzzle solving A* every day. And trying this, a lot of the actual code turns out to be stepping stones, or refactorings.
So the other day, i figured out exactly how to fix the root motion on the mixamo rigs i had downloaded. They would be in FBX form with motion baked into the object, not the root bone. And what was holding me back was the fact that the object rotations were in ‘XYZ Euler’ form, but the bones were in Quaternion form. The keyframes couldnt be copied and pasted directly.
So I decided asking google for a script to fix the root bone was a complicated question, and instead asked for ‘blender python how to change Euler to Quaternion’, then i followed up with ‘Blender Python how to access/set keyframes from an object’, and then ‘bone’ or words to similar effect, and then i started getting results in the form of usable scripts …
Im sure this saves time just as google searches did in the old days.
It turns out the root can be fixed by copying the keyframes from the root object (named Human.rig) making sure the root bone is unselected, then deleting them, and selecting the keyframe on the root, deleting, then in the bone properties switch rotation to XYZ Euler, then go to frame 1 with root selected, insert a keyframe, then just paste over.
I could still try to script it, but there are a bunch of snag cases where the root goes under the floor, and the manual method has worked fine.
1 Like
Another thought about all this I just had. A lot of the companies using AI to replace workers are going to find in the future that they predicated their own downfall. Because at some point, LLMs or a descendant of them will become god enough to replace their entire company. And it’s not going to be as good, but it’ll be “good enough” to where people will turn to LLMs because they’re “cheaper” or “less hassle” and those companies will not be able to retain enough business to stay in business.
Reminds me of Friedrich Niemöller’s poem about Nazi Germany.
First LLMs replaced the data entry people, and I said nothing.
Then LLMs replaced the junior developers, and I said nothing.
Then LLMs replaced the artists, and I said nothing.
Then the LLMs replaced me, and there was no one left to speak out.
4 Likes
Ouf, you reached a new level of bleak there 
Just to draw a very important line here: we are not talking about systematic destruction of lives and I guess your comment is intended as a satirical oversaturation.
This is probably more of an ignorant collateral driven by greed and blind feeling of progress. Another point in which it is kind of related to the dotcom crash, as well as any other technology driven crash in history. One would wish we would have learned from the times before, but here we go again.
Also, “AI”, as non-AI they are at this point, luckily don’t have intention.
2 Likes
People tend to forget that what we call AI at the moment not only is wasteful in precious energy ressources (electricity consumption bigger than whole countries’) but is dependent on exploitative, enshittified human labour in (mostly) developing economies to do the labelling and tagging jobs on the data that will be used by LLMs to do their “magic". Pure greed : destroy good job opportunities in the first world to go on exploiting the developing world with piece work (data annotations jobs are the “better" paid tip of the iceberg).
It’s a bubble of greed and deceit.
4 Likes
Absolutely, but perhaps the AI can be used to break down social barriers.
One of the best programming paradigms ..
Be agile, adapt to change.
1 Like
Drink the kool-aid. 
1 Like
If your interested in at a modern look comparing the current theoretical AI bubble vs the 2000 dot com bubble there is this video by coffeezilla on his second channel
He explains how the dot com bubble happens specifically at starting 5:05, main explanation at 5:53
2 Likes