Ah yes, the famous "trickle down" theory, which you may know from its smash hit "trickle down economics"! Surely this time it'll be different, right?
Ah yes, the famous "trickle down" theory, which you may know from its smash hit "trickle down economics"! Surely this time it'll be different, right?
> Ah yes, the famous "trickle down" theory
While trickle down theory is dubious when it comes to economics, it's clear that it is valid when it comes to technology. Tech trickles down.
Ah yes, like my RAM, SSDs, and GPUs are trickling down right now? You're absolutely right! I love a trickle of PC parts!
This is a temporary anomaly due to the datacenter buildout, which will not last forever. Historically, tech has always trickled down. Sooner or later, hardware will again become cheap.
"Temporary" is a few months. We're looking at many multiple years before parts even _begin_ to decline in price, if even. How much will we as consumers lose in the time it takes for this to correct? This is not "temporary", this is a decimation of an entire product sector for the foreseeable future. There is no historical precedent for this. Companies are not and will not slow down gobbling up literally every tech part they can. Every year, training and running AI will demand more resources over and over and we will continue to be charged increasing amounts for everything because everything costs more either directly or upstream of us. It is a literal cancer.
> Every year, training and running AI will demand more resources over and over and we will continue to be charged increasing amounts for everything because everything costs more either directly or upstream of us.
Unless we are really going to hit the Singularity and create Godlike superintelligence that improves beyond all limits, sooner or later the models will stop improving, no matter how much compute we throw at them. After that happens, the best models will be etched into model-specific ASICs with weights fixed in mask ROM (like Taalas is already doing), which will result in massive, order of magnitude inference efficiency increase. Couple this with the training datacenters suddenly becoming available for inference, and the cost of compute will plummet.
Yes, it might take a few years, but the hardware will be cheap again. And if the result of the whole process is human-level AI, then it will be worth it in the end.
I don't think you understand the current state of capitalism very well. Absolutely nothing goes down in price once companies know people will still buy an item at a higher price. We will never see hardware at the previous affordable prices again. Sure it might go down somewhat, but going down from 500% of previous prices to 400% of previous prices is not suddenly "affordable". And that is exactly what will happen. Models will not suddenly stop improving and their consumption of hardware will not stop. The corporations won't allow it. We have too many companies in an arms race against each other. They literally cannot stop or they die.
How can you say tickle down economics doesn't work? The US has made it easier for businesses to grow and the result is the highest standard of living ib tje world.
I think it's hard to quantify "highest standard of living", but for the attempts at it (e.g. HDI, etc), the US is not number one.
Trickle-down economics specifically refers to economic policy that favors the wealthy and corporations with the argument that this will lead to wealth trickling down to the economically less fortunate. Examples include the Reagan and Trump tax cuts.
This has been studied somewhat extensively (https://academic.oup.com/ser/article/20/2/539/6500315?login=... is one example), and it’s been shown over and over that this does in fact not work and leads to larger wealth disparity, across many countries where it has been attempted.
It’s possible you’re working off some other definition of the term based on your comment but that’s the generally accepted usage.
The person you replied to was the second person I've seen on HN this week try and defend trickle-down economics explicitly in more or less the exact same terms and justifications. What is going on with that? Did Peter Thiel tweet some pro trickle down dreck recently that these people latched on to?
You should see the comments sections on NYT articles, lately. It was like someone flipped a switch about two weeks ago.
It would be neat if someone did sweeping sentiment analysis on them, just to see it graphed, but I’m sure gathering the data would be prohibited.
It’s all bots anyway at this point.
Hate to break it to you, but the US doesn’t even break the top 10 in a lot of reports.
The US alone!