Also a major contributing factor to second (and third) order effects. Humans, groups, even societies all respond to changes - and if you cannot anticipate what those changes might be, then you are out of luck.
(You will never get all them right. You will never even be able to list what their entirety will be. But you have to be able to predict the order of magnitude of a few of them.)
As much as the site is an incredible outlet for absurd creativity, some of the creations would actually work as small batch orders. The octopus hoodie is a great example, and I would not be surprised if there were people willing to get different variations of it. (Lovecraftian flavour, anyone?)
OP: well done, you have unleashed on this world a toy more addictive than a cocaine enema.
>you have unleashed on this world a toy more addictive than a cocaine enema.
No offense to you meant, but I wonder in general where the need for this kind of hyperbolic phrasing comes from. As it seems to be everywhere on the internet.
It was meant as a compliment, and it was not intended as hyperbolic. But since you ask...
In this case I thought it would fit with the already absurd tones exhibited in the thread. More generally, the technique is not "hyperbolic phrasing" as much as deploying a comedic angle. Comedians (especially oneliner and short-form comics) often seek ways to emphasise a visual image. The more vivid the mental imagery, that much more effective the double punch of the words and the internal visual hit.
The same technique is also occasionally used by some of the most effective tech talks; if you manage to combine a factually correct detail distillation with a punchline that invokes a strong and somewhat controversial mental image, that has a high likelyhood of being remembered.
Just wait until the foundational models are all fed with increasing amounts of ads. After all, it will be the one remaining source of ongoing, if not human generated, at least human curated content. Ad exchanges will offer firehoses for a hefty fee, and the advertisers themselves will pay the exchanges extra to push their ads more frequently in there to gain a higher share of repetition.
The AI companies in turn will hoover in the deluge because they need something new to train their models with, embedding the ad copy deeply into the model itself.
Your local AI of the future will be just as ad-riddled.
> Just wait until the foundational models are all fed with increasing amounts of ads.
Which foundational models?
Not all model providers are in the ad business and while the chances of building a supercomputer in your basement to train such a model are zero, some of the companies that build such models aren't exactly huge. Mistral for example is (according to Wikipedia) 150 people - this means that a company that can make their own model from scratch doesn't need to be some giant corporation. Which in turn means that it is possible for new companies to pop up, if there is a need for them - in this case, if some models become ad-infested, chances are other models will use their ad-free status as a feature.
And this is assuming only companies make such models. But some days ago i was reading here in HN about a new foundational model being trained by ETH Zurich and somehow i doubt a public university will inject ads in it.
Yep, 'dataset poisoning' is what I like to call it. If all newly-trained models mostly reflect the Internet, then the strategy is to shove your ad copy into such a large fraction of the Internet that models will accept it as true. Like today's SEO slop, but turned up to 11, since it's not even aiming for human clicks.
> Why is Amazon gobbling up failing studios on the cheap?
For their back catalogue. The audiences are after familiarity ("more of the same") and nostalgia. The studios in turn are terrified of taking risks, which is why nearly everything they release has been a sequel, prequel, reboot, or in-universe spinoff for more than 15 years. Buying up a studio gives access and control over their massive quantity of pre-chewed dough to feed their cookie cutter productions.
In this Amazon behaves like a PE entity: buy up for cheap, roll up what they can, reduce quality, and milk for as much money as possible before tossing the empty (& possibly toxic) husk aside.
> Why did Amazon, Apple, and Netflix offshore production to the UK, Eastern Europe, and Asia when the US has infrastructure and subsidies?
They have been following what the studios have been doing for decades. Even in 1990's a good chunk of US productions were often filmed in Canada. (To lower production costs, of course.) UK was known for some of their film studios even before then, so made a good early target when Canada started to become too costly and you needed to move elsewhere.
You can see where this is going. We're already seeing more productions filmed and located in/near China[ß]. As the costs there will eventually creep up too, expect to see locations shift to places like Indonesia, Pakistan and South-Eastern Africa.
ß: Part of China's visibility and increased footprint on films is due to a clear political drive. The CCP has set up structures where they strongly encourage their industry arm to fund film and TV/streaming production elsewhere in the world, buying up influence and dictating how Chinese get portrayed in the scripts.
For a brief period of time, binary configs[0] were a thing. In mobile world only, but still. It wasn't that people generally wanted them, but because random seek I/O latency on early mobile devices (and especially on their eMMC storage devices) was atrocious.
Opening up tens or hundreds of XML config files for resync was disgustingly slow. I've developed software on Maemo and Scratchbox; the I/O wait for on-device config changes was a real problem. So of course someone came up with a modified concept of Windows registry - a single, binary format config storage, with a suitably "easy" API. As a result you'd sacrifice write/update latency for the cases where you wanted to modify configurations and gain a much improved read/refresh latency when reading them up.
Of course that all broke down when reading a single config block required to read the entire freaking binary dump and the config storage itself was bigger than the block device cache. Turns out that if you give app developers a supposedly easy and low-friction mechanism to store app configs, their respective PMs would go wild and demand that everything is configurable. Multiply by tens, even low hundreds of apps, each registering an idle-loop callback to re-read their configs to guarantee they would always have the correct settings ready. A system intended to improve config load/read times ended up generating an increased demand for already constrained read I/O.
Consider the second order effects of building all those data centers.
The GPU hardware rots and becomes obsolete in a matter of years, but the national infrastructure required to support the physical sites isn't going away. Things such as...
- improved power distribution networks
- logistics arrangements to build and support the DC sites
- lots and lots of new fibre interconnects to support the massive bandwidth needs
- hopefully: better power delivery planning laws
- plumbing infrastructure, because all that hardware requires cooling
Some of the DC sites will be decommissioned from their initial use, but given the physical security requirements, might morph into handy higher-security industrial facilities with only small repurposing. Such reuse cases would especially benefit from improved logistics (see above).
That's fine, but the non-GPU infrastructure represents half of the cost of the datacenter. The physical GPUs and compute represents the other half and will rather rapidly depreciate.
This is not as hardy as fiberoptic communication lines used to build the internet, or railroad lines used to build transportation infrastructure.
I'll be honest: this list sounds like grasping at straws to find something positive in this fiasco and doesn't come close to building something lasting and valuable for USA.
The cure for the spam problem is to turn the incentives around in the patent examination process. Make it highly adversarial from the go and reward the patent office staff for rejecting applications, with solid argument trails. A good part of that would be to have the applicants submit their researched potential prior art and PROVE that their patent is actually novel. (If they haven't submitted a clearly discoverable piece of prior art and the examiner finds that out, that's an immediate rejection. With extreme prejudice, and preferably multiplied rejection award.)
It would still be gameable (everything is), but it would certainly curb the flood of copycat "X-but-in-domain-Y" patents once the pool of prior art used to reject crummy patents becomes better known and established. The additional pool of rejected applications then also feeds into the prior art foundations.
I know you're joking, but this doesn't feel too far off the mark in this world of late-stage capitalism run amok. Give it another 15 years and the bleeding edge[!] insurance companies are likely employing agents to go after clients who have become a net drain on their P&L.
The agents probably won't be doing that "themselves", but instead will be offering bounties (think: contracts) on suitably well hidden assassination markets. After all, as a machine AI cannot be held accountable for what is essentially a management decision.
I'm personally still waiting for the first country to go full Running Man to solve their prison overcrowding issues, and in addition to entertainment licensing deals also offer state-sanctioned gambling options to get a second bite.
United Healthcare already did this with their AI powered death panels that would bulk deny claims, even if they later allowed them, the delta-t causes bonus subscriber terminations.
> Is there something materially different here with the internet?
Yes. There are effectively no hard copies. It is possible to change the historical record of any non-printed material to suit your particular needs like never before.
You can think of this as a world beyond Orwell's or Bradbury's wildest nightmares.
On the other hand, it is now easier than ever to make copies of materials that we see. Famously, the internet never forgets, and even the smallest mistakes or slip-ups are retained in perpetuity, as long as someone is interested enough in keeping hold of the original copies. And there are a lot of organisations that are very interested in keeping hold of original copies.
I would argue that the opposite is true: it is now harder than ever to change the historical record, which is why we now talk about hypernormality and post-truth, where even if there is evidence for something, people will still lie and claim the opposite and be believed. We live with an abundance of evidence, and yet the Orwellian ability for people in charge to tell you one thing one day, and another thing the next, has never been stronger.
And I think you're again making the mistake of thinking of the printing press as a device for printing books or other materials designed to be long-lasting and valuable. In practice, the printing press brought about a revolution of flyers and pamphlets - ephemeral documents that were distributed one day and then abandoned the next. These things should change freely, and many never entered the historical record at all.
This is why I think archive sites will be attacked by the powers that would like information to disappear when they want it to.
Perhaps they'll use a warped interpretation of copyright law to do it, or maybe something even more draconian like censorship laws with a punishment for publishing banned information.
Could they do it, technically? Not unless they controlled the entire world's networks, including those of countries with competing aims. Would that stop them from trying? As we've seen with the endless attacks on end to end encryption, I'm sure they'd give it a shot.
I don't think it's possible, though. Or at least, I think it's harder now than ever before. The internet isn't completely decentralised, but it's at least spread out enough that it's seriously difficult to shut down any one part of it, at least without being willing to take some serious authoritarian measures. Look at how difficult it's been for the most influential media companies in the world to fight piracy, for example.
I genuinely think our society is one of the most censorship-resistant societies in history. This comes with its own problems (how do you deal with media that genuinely is harmful, like calls to violence or plots to abuse children?) but I think this is the tradeoff that one has to make when dealing with censorship and liberty. The more you make it difficult for the authorities to shut down good speech, the more difficult it becomes to protect against harmful speech.
> On the other hand, it is now easier than ever to make copies of materials that we see.
Is it? I feel it's the other way around. For example, just 10 years ago, there were no apps that forbid me from taking screenshots. Copying CDs and DVDs was comparably easy, even for normal folks. How do I copy a Netflix episode again?
I meant "now" in that sentence to refer to the internet era in general, in comparison with other technological leaps. But still, these protections are usually very limited, and fairly easy to circumvent. Most people I know might not be able to convince their laptop to let them screenshot Netflix, but they do normally know how to find pirated copies of the TV show they want to watch. Paying for the convenience of Netflix might still be worth it for them, but the ability to step beyond that should that convenience disappear is definitely there.
I agree with that first sentence, but I think the trend matters more than the average over a few decades.
Also, you said "On the other hand, it is now easier than ever to make copies of materials that we see." - now you seem to be talking about finding a copy. But those are two very different things.
I'd like to think I'm not. In the past, at least there was an original "hard copy" - and any regime wanting to rewrite history would have to meticulously either eradicate any prints, or - as happened in the Eastern Bloc - they would have to physically rewrite history. Pages from books were lifted out, edited to suit the needs of the narrative, and then meticulously put back.
With online-only records any hard copies will be incidental. The source-of-truth for any record has always been online, and can be retroactively edited with much less fanfare. Incidentally, it will also be much easier to flood the world with the updated narrative.
Hell, we have Musk publicly advocating to edit old online material to suit the new, "more desirable" narrative.
(You will never get all them right. You will never even be able to list what their entirety will be. But you have to be able to predict the order of magnitude of a few of them.)
reply