Gormo

This is a basic concept of Public Choice Theory.

People worrying about money in politics are overemphasizing the instrumentality of money, rather than the underlying structural incentives that tend toward special interests dominating complex issues.

If money were removed from the equation, the concentrations of power, motivations to influence it, and asymmetries in those motivations will all continue to exist; political influence will just be exercised through other means, and some particular faction or class of people will still be at the sweet spot of having the greatest competence and the greatest motivation to apply those means.

A lot of people are reluctant to acknowledge it, but this is inherent to politics itself, and there is no straightforward way to fix it, other than to limit the reach of political power.

If you don't have the solutions, how do you verify Claude's work?

> Yeah, that's not going to solve twitter, facebook, instagram, etc.

It sounds like that would precisely be a solution to the problems of using these platforms.

We were in the "Chase the Chuck Wagon" phase of the TV and movie industry long before AI was a factor.

Maybe they could divide movies into a series of standalone segments that tell the story one chunk at a time, and then begin each segment with a short recap of what happened in the previous segments. They could even release the segments sequentially over time, and keep people coming back to it over the course of weeks or months.

I don't think we're disagreeing.

The point at which new changes are more likely to degrade an application that's already excellent for its established use cases than they are to improve it further is the point at which we should consider the software "finished".

There may be lots of theoretically plausible improvements that could be made, but if the constraints of actual practice make pursuing those ideas a net detriment, that's probably a good place to stop.

I think AI may actually make this simultaneously better and worse: it will eventually be really easy to hit all of the functional and performance targets early on, so "finished" software will at some point be much better and closer to true completion, but the mentality that drives people to keep fiddling with it instead of moving on to a new project will be just as prevalent, and the post-"finished" releases will likely be even worse.

> Most consumer software in developed under the assumption that features only have a cost when being built. Their maintenance cost, and the combinatorial complexity each new feature adds, is barely considered.

I suppose that's true, but I don't see what it has to do with UI design from a usability standpoint.

> But this doesn't mean the software couldn't still be improved, just that there isn't the resource available to do so.

It's a push and pull from both sides, isn't it? The accretion of process complexity among the developers may increase over time, but so too does the completion of core features and solution of all of the low-hanging-fruit development tasks.

Pretty much every GUI application I've ever used seems to hit a point where the original purpose for which the software was devised is satisfied for essentially all common use cases, but the incentives of the dev team are to keep going, so they move more and more toward gimmicky features, unnecessary bells and whistles, visual changes to the UI, etc. because any real functional improvement is a big investment, but they still need to justify a new release.

UI churn is one of the most common failure modalities here: when you've already picked all of the low hanging fruit in functionality and performance, major UI changes can create the appearance of improvement and innovation without actually having to do any of the hard work to achieve real improvement.

So when I see unnecessary UI changes to an application that still basically does the same things in the same way, at this point I'm conditioned to assume that that application has hit the point of diminishing returns for actual functionality, and the development incentives are now going to be driving more and more churn. And that's when I consider the software "finished".

> because powerful AI requires tons of compute for training

What explains the prevalence of models being trained by small organizations, or even lone individuals, on relatively affordable hardware, then?

> so training a frontier model requires billions of dollars

The term "frontier" here almost makes this a tautology -- "the kinds of models that currently cost billions of dollars to train are only being trained by people who have billions of dollars". But things are moving so fast that the frontier of the immediate moment will be well-trodden ground a few months from now.

Ground-up models are emerging at an accelerating pace, and the FOSS community is producing distillations of proprietary models rapidly.

> That's never the case for any consumer app, the potential for improvement in UI is practically infinite.

If that's true, how do you explain the observable pattern of many consumer applications degrading in UI functionality and coherence after a certain point?

> But also, just use the old version if it's finished, you don't need to upgrade?

Exactly. I reverted to 0.12.

I've been using it for years and was surprised to see a new version after a dacade.

After installing it, I was surprised to see that the immediate changes most noticeable to me were that the input box was unnecessarily doubled in size, enclosed in an unnecessary border, and surrounded by unnecessary dead space; and that there's now marked input lag that never existed in the original version.

I never had a single problem with 0.12 over the past ten years, and use it nearly every day -- no bugs, no missing features.

I appreciate the desire to improve and innovate, but you know, sometimes, software really is finished.

> So, the course of AI is currently in the hands of ~10 billionaires.

Is it? It seems to be spreading around everywhere. There are tens of millions of users and hundreds of thousands of organizations on HuggingFace. How are the few of people running the merely most publicity-focused companies in the space in control of the whole ecosystem?

Yes, another person in the thread found that already. The provisions in 302a don't originate from either of the statutes you cited, but rather from Public Law 90-379, enacted by Congress in 1968: see https://www.congress.gov/bill/90th-congress/house-bill/14910....

It sounds like it will be up to the courts in this case to determine if these specific FCC regulations are actually made pursuant to the purpose of that section, and whether operating a marketplace within which others are offering radio equipment for sale can be interpreted as you yourself offering that equipment for sale.

> Because the things you have in the fridge change.

Are the contents of the fridge also being adjusted via remote management tools?

Yes, but the point is that the motivation for taking these actions isn't for them to be effective, but to satisfy performative requirements imposed by external parties.

No, it doesn't appear that the 2019 act you referenced, nor the 1934 Communications Act, specifically relates to this.

> It's not obvious the subsidization will end. Loss leaders are a thing.

Loss leaders are a thing so long as they actually do lead somewhere. If your primary value proposition is conveyed by the loss leader itself, where do you go from there?

It might be an idea that Marx incorporated into some of his thinking, but it's not a "Marxist idea" in the sense that any variation on it must carry with it all other specifically Marxist precepts.

Marxists themselves were historically quite guilt of the same error, in that they assigned objective value to outputs and prioritizing raw output over subjective utility.

There's also the underlying problem of trying to evaluate "objectively right/wrong examples" in the context of decisions that depend on mutable subjective variables.

Chess mostly boils down to a mathematical optimization problem within a fixed, predetermined rulset, with predefined success and failure conditions.

Geopolitical decisions involve pursuing outcomes in situations involving far more variables, often changing in realtime due to factors outside the "game board", where the rules are complex, mutable, and not necessarily known in advance, and where the success and failure conditions are subjective and may themselves change during the course of the scenario.

Geopolitical situations are Chess, Nomic, and Werewolf all rolled into one. How good is AI at winning all of these games?

I appreciate the reference to an actual, applicable statute. Based on that, it sounds like whether eBay acting as a marketplace is actually the party offering equipment for sale, rather than merely providing a venue for others to offer it for sale, is something that would have to be figured out by the courts, as would the question of whether the specific regulations the FCC is enforcing actually are pursuant to that section.

> If you believe the FCC or Congress misapplied the interstate commerce authority in the constitution or prior case law

I'm looking for the specific statute in which Congress exercised its interstate commerce authority to actually assign this power to the FCC.

> this SCOTUS has expressed interest in reverting the interstate commerce authority to its pre-New Deal interpretations, which would disintegrate the administrative state of the last 100 years

The the post-New-Deal interpretation of the interstate commerce authority still relies on understanding that the power to regulate interstate commerce belongs to Congress, such that administrative agencies only have those powers specifically granted to them by statute.

So regardless of this, the FCC still doesn't have direct interstate commerce authority under the constitution, and can only exercise such regulatory authority as has been assigned to them by by Congress.

Whether or not those statutes will continue to be recognized by the courts as a legitimate exercise of commerce clause powers is another matter entirely.

I looked up the Secure and Trusted Communications Networks Act of 2019 (https://www.govinfo.gov/content/pkg/PLAW-116publ124/pdf/PLAW...), and it appears to be the act under which products from certain manufacturers were excluded from being purchased with federal funds. It seems to make no mention of either regulating domestic marketplaces or giving any powers to the FCC. Are you sure you're citing the correct statute?

"PIE" in this sense being an abbreviation for "Proto Indo-European".

Try "hampster".

One of the things I notice is that a lot of these aren't generic decorative images, but are unique to a specific website. It'd be nice if this archive identified the point of origin, and linked back to a snapshot (or to neocities.org) so we could see them in their original context.

What statute gives authority to the FCC to "make up rules for marketplaces"? My understanding is that they only have authority over actual broadcasting.

"Models doing rogue things" is just another way of saying "complex system operating out of control limits".

Software engineering has until now dealt with deterministic systems that operate on Boolean logic, so dealing with stochastic processes involving unaccounted-for variables has never before been something we've had to deal with. But over in the real world, people dealing with physical, mechanical, operational, or even social systems have had to deal with the failure modalities of stochastic systems forever.

Maybe when working with LLMs as functional tools, it'd be wise to bring some of the wisdom from operations research, quality engineering, etc. into play.

This is exactly correct. I don't understand all of the arguments that attempt to special-case LLMs. LLMs are just a different way of giving instructions to a computer, and one that we know is based on stochastic models and therefore has a higher likelihood of malfunctioning than conventional software.

But an LLM behaving in unintended ways is functionally no different from a conventional program behaving in unintended ways due to bugs or design flaws.

I think the narratives that anthropomorphize AI are doing everyone a major disservice in leading some people to think that the standard rules and expectations somehow don't apply when it's involved. But it's still just software, and you're still responsible for the results of software that you deploy and operate.

> But also, the agents absolutely went rogue.

If a car crashes into a building because its owner left it in neutral and failed to engage the parking brake, did the car "go rogue", or was it just being operated negligently?

Where are you getting your info from? Wholesale markups are typically 100%, i.e. resellers generally do capture ~50% of the final retail price of merchandise.

If Amazon is charging that while also offering advertising/product discovery services that traditional distributors wouldn't offer, it sounds like a bargain based on what you're describing.