The idea is that you run a workload at a model provider, that might cheat on you by altering the model they offer, right? So how does this help? If the provider wants to cheat (they apparently do), wouldn't they be able to swap the modelwrap container, or maybe even do some shenanigans with the filesystem?
I am ignorant about this ecosystem, so I might be missing something obvious.
I wish everyone a nice party, but the problem isn't Trump. Its what behind him: the ideologues, the power brokers and their networks, the 0.001%. Plus the masses having been bathing in culture wars for years.
Trump is just good for circus, I would say the GOP can call themselves really lucky with him. His job is to successfully capture media attention, keeping what enables him out of the spotlight. He lacks all qualities, except that one ability to grab the mass media by their pussy. New craziness every day makes good headlines.
Problem is that his enablers are not aligned on all core issues. Yes, you have got the Heritage Foundation which mainly wants to go back to the gilded age with a vast christian lower class. But you also have a circle of people who believe that crashing the US, including the dollar, enables them to build a US like they want. Its a weird coalition of billionaires predating on the millionaires, grifters, christian nationalists, Neo-nazi's like Miller, tech-accelerationists etc.
You should fear the day when Trump isn't needed anymore. MAGA is Trump. GOP will have to shift up ideological gear after him, and it won't be as nice as Trump. Even if internal war breaks out in the GOP, it is too early for a party.
> people who believe that crashing the US, including the dollar, enables them to build a US like they want
Yes, it's strange how dumb some rich/succesful people are. As I understand it, no civilization ever has done such a thing. If a civilization and its institutions crash, it remains failed/dysfunctional for a very long time. The only way to improve society is in small steps.
I hope the people who finance this all will wake up to the reality that it may well cost them everything, too.
It's not strange; they can just afford to weed out the people who say "no" from their lives. Everyone around them is either in the same situation, or depends on them for their cushy livelihood.
Not having to hear "no" for decades breaks brains.
You're right. Trump does an excellent Zaphod Beeblebrox. He distracts from power, and I get that, but he's still a piece of crap, and a lot of people have died from his fumbling, bumbling, inept, failing upwards solely due to the fact that people associate him with having money and power, even though he's an tryannical, ineffectual, foppish, childraping manboy.
Because we hn people are used to reduce the world to a set of technical parameters. I am not intending to blame or shame anyone here, but to take it more broadly, the discussion around Doge showcased many such problems that arise from unawareness about the limits of our approach: context blindness, taking narratives at face value, narrow focus on technicalities, no consideration for ethics etc.
Tech people need to reduce complexity to make it computable, that's our job. Our strong points are the weak points too. Again: no blame or shame. Just wanted to point out we are susceptible in these matters.
I think you have it backwards. Regulation is what enables monopolies. There is no monopoly for any of the major industries like cow herding or cellular telephone service in Somalia despite almost no effective regulation. There is not even a monopoly of pirates despite them willing to use violence to try and enforce a monopoly.
If you look at the history of the US, for instance, railroad regulation was brought forth largely by the railroads because they found it impossible to form a cartel to keep up prices (due to "secret" discounting) so instead they created regulation that outlawed the kind of discounting that breaks cartels apart. A similar thing happened in banking where the banks asked for a central bank to cartelize the interest ranks to stabilize their oligopoly. And the same in pharma industry -- big pharma loves high regulatory barriers because it keeps competitors out.
A large portion of the regulation in the US was brought about as regulatory capture by corporations to increase the monopolizing effects and destroy the free market.
That's patently false. AT&T was a monopoly and they were broken up by antitrust regulation. The absolute most you can say is that some regulations enable monopoly. I contend that we simply should pass the good kind of regulations instead.
Monopoly is enabled by market forces such as economies of scale. Monopolization is a natural market process which happens on its own unless it is actively prevented.
> big pharma loves high regulatory barriers because it keeps competitors out
The FDA, for all the flaws of its current incarnation, is the archetype of necessary regulation. Pre-FDA, the free market did nothing whatsoever to prevent nauseating practices like the adulteration of milk with powdered plaster, lead, and cow brains. The history there is fun but quite gross.
> Somalia
What is notable about Somalia is not its lack of regulation, but the fact that it is perhaps THE least stable country on the planet. It is not the basis for any useful comparison here.
>That's patently false. AT&T was a monopoly and they were broken up by antitrust regulation.
This is patently false in the context of the reply you have made -- after the invention of the telephone more and more and eventually hundreds of telephone services popped up. Then in 1918 (circa WWI), the government effectively quasi-nationalized AT&T by controlling it via a commission and the postmaster general and then AT&T leveraged politicians to create "universal telephone service" provided by AT&T and regulate competitors out of the market while using regulatory capture to use commissions to regulate rates, effectively creating a cartel that drove competitors out of business via regulation.
the whole idea of a "natural market process" here is absolute and utter hogwash. The majority of the market was AT&T competitor up until the regulators stepped in and turned it into an unnatural monopoly enforced by regulatory capture.
>The FDA, for all the flaws of its current incarnation, is the archetype of necessary regulation. Pre-FDA, the free market did nothing whatsoever to prevent nauseating practices like the adulteration of milk with powdered plaster, lead, and cow brains. The history there is fun but quite gross.
You're now arguing why we need regulation rather than whether they create monopolies or not. I see this as a complete red herring, although an interesting topic, that there are some counterpoints to.
> What is notable about Somalia is not its lack of regulation, but the fact that it is perhaps THE least stable country on the planet. It is not the basis for any useful comparison here.
What is notable is that the whole thesis is without regulation it turns into this monopolized hellscape and every inspection of that theory turns out to be false, and sometimes even the opposite.
Anyway, what regulation is responsible for Walmart and Amazon putting local retailers out of business?
> the government effectively quasi-nationalized AT&T
After a big merger put AT&T in charge of the majority of telephone lines in the US, the company used its control over infrastructure to drive its competitors out of business and increase its portfolio. The Justice Department tried to break up AT&T but failed; it was in the settlement of this case that AT&T was first federally regulated in 1913. Yes, AT&T's monopoly grew between 1913 and 1982, but your causality is backwards. They regulated it because it was already a monopoly.
>Anyway, what regulation is responsible for Walmart and Amazon putting local retailers out of business?
Walmart + Amazon combined are only ~16% of the retail business. They're not monopolies. The fact they put a small minority of businesses out of business does not mean they're a monopoly. This is likely part efficiencies and also part regulatory capture via the insane zoning/building regulations in this country and tax breaks that can favor large corporations.
> After a big merger put AT&T in charge of the majority of telephone lines in the US, the company used its control over infrastructure to drive its competitors out of business and increase its portfolio. The Justice Department tried to break up AT&T but failed; it was in the settlement of this case that AT&T was first federally regulated in 1913. Yes, AT&T's monopoly grew between 1913 and 1982, but your causality is backwards. They regulated it because it was already a monopoly.
... It was not already a monopoly. Hundreds of phone companies emerged and by shortly before your noted date of "regulation" those competitors held the majority market share. It became a "monopoly" after the government literally quasi-nationalized them (AT&T) to the point the fucking Postmaster General was basically in charge of it, they became intertwined with regulators, and then the drive for "universal telephone service" and regulatory commissions ensured the regulatory compliance pushed exactly into AT&Ts business model. AT&T intertwined lawmakers even brought in economics quacks to talk up natural monopolies to argue for policies to create the regulations that made AT&T a monopoly. So you have it backwards -- the regulated it from a minority market holder to an unnatural monopoly and lawmakers created this monopoly under the auspices they essentially needed to legislate a "natural" (misnomer) monopoly into existence.
> Walmart + Amazon combined are only ~16% of the retail business.
of ALL RETAIL? That includes groceries! That's huge! Anyway, Amazon is >40% of e-commerce & Walmart is >10%. Together they control more than half of all online commerce. That's definitely monopolistic.
> It was not already a monopoly. […] shortly before your noted date of "regulation" those competitors held the majority market share
From Wikipedia: "AT&T controlled over 80% of the U.S. telephone system market by 1907 and Theodore Newton Vail rejoined the company as its President. Vail negotiated with competitors, charging them fees for connecting to AT&T's long-distance network. These practices led the Justice Department to attempt to breakup AT&T, but a settlement was reached through the Kingsbury Commitment on December 13, 1913. It brought federal oversight into AT&T and led its Bell System monopoly to become federally regulated."
They had an 80% market share pre-regulation. Yes, it was already a monopoly.
>They had an 80% market share pre-regulation. Yes, it was already a monopoly.
Absolute and utter hogwash -- a straight up lie. You must be using a very liberal version of 'control.'
After seventeen years of monopoly*, the United States had a limited telephone system of 270,000 phones concentrated in the centers of the cities, with service generally unavailable in the outlying areas. After thirteen years of competition**, the United States had an extensive system of six million telephones, almost evenly divided between Bell and the independents, with service available practically anywhere in the country.[1]
* My note, under the government imposed patent monopoly period -- Bell's patent expired in the 1894 which started the "years of competition." **13 years of competition marks 1907.
That is, by 1907, the market had dropped from a patent imposed monopoly to half-and-half, and getting shredded further by competition (that is until regulation in 1913, when AT&T started to pick up market share again). This 80% quote is total fiction unless you're using some weasel version of 'control.' The monopolies were all government imposed -- first by the patent and then later by regulation ('universal telephone service' reguluation and commissions and franchises, also ~nationalization and government intervention circa WWI).
Also of interesting note -- to look at the ownership of telephones before and after the Kingsbury committment. They had been falling off a cliff after the patent expired, but then ramped up at pretty much the same time as the Kingsbury Commitment (minima at 1910, with what looks to be 5 year granularity).[2]
>of ALL RETAIL? That includes groceries! That's huge! Anyway, Amazon is >40% of e-commerce & Walmart is >10%. Together they control more than half of all online commerce. That's definitely monopolistic.
And 100% of business with an Amazon logo on it! Amazon has 40% of e-commerce, walmart has 6%, even together they are a minority. Even with all the efficiencies of Amazon logistics they still together can't break half of the market. And even if they did, their margins are so razor thin that they could not engage in monopolistic behavior, as the second they raise prices they can again be eaten alive by the other 54% or one another.
This is not a personal opinion of mine, it is pretty much established science. I think only think-tank backed sources would claim the opposite.
One should understand the phenomenon as a common pattern of dynamics in unregulated markets. Not every snapshot will showcase an end state of monopolist dominated markets.
You bring up a valid point though. Regulatory capture is a indeed a weapon in the hands of anti-competitive players to prevent incumbents. Good policy usually applies differently to different strata: the small players are exempt from certain rules, or have to deal with less stringent ones than big players do, to prevent killing the market. At the risk of sounding like an llm: it is not just about policy, it is about good policy.
This is a crude but misguided attempt to bypass what I say which was "effective regulation." FGS only controls a minority of Somalia -- most of it is controlled by Somaliland, Puntland, Al-Shabaab, directly by xeer (customary) law, or only FGS hanging on by a thread.
Even in cases where FGS is in control -- even then xeer law on property rights and law often supersede written law of FGS. For instance, on occasion Somali population has straight up killed FGS soldiers/police and Somalian government has deferred to xeer courts and said "welp that is fine." Xeer law in particular has a very liberal free-market view on inter-familial entrepreneurism (although with a lot of intrafamilial and tribal dues which hinder it in practice) which is at odds with what outside law has tried to impose upon it.
NCA has rather limited influence in somalia, and definitely not of the sort that could break up a monopoly if one existed.
The food industry is filled with regional monopolies.
> cellular telephone service in Somalia
Ah yes, excellent example, all you have to do is completely destabilize your nation and you too can have free market capitalism.
Investors love monopolies, they fix prices and profits so their investments are not at risk. Investors hate too much competition, it lowers profits and puts their investments at risk.
Free markets need investors. Investors hate free markets. I hope you see the problem here.
This sounds like almost the best business environment for criminals.
"I am sorry judge, yes, it could be that we are involved in crime, but we have been too busy counting billions of dollars each year. As you might understand, businesses are not part of society, they should only be judged on their shareholder value. We reap the profits, society pays for the collateral damage, that's only fair."
Yes, you mentioned leeway. That would only make sense in the context of an entity understanding it's role. It does like in the way above.
> Some of the rise in billionaire spending can be explained by growth in the number of billionaires — but not much. The number of U.S. billionaires rose 85 percent between 2010 and 2023, from 404 to 748. But billionaires’ share of political contributions rose by 1700 percent.
> In short, we are in the midst of an unprecedented power grab by America’s oligarchs. This power grab is arguably the most important fact about contemporary U.S. politics. In many ways MAGA is just a symptom.
> What lies behind this power grab? An extraordinary concentration of wealth at the very top. (..)
> That said, even billionaires care about more than their own personal wealth. Unfortunately, their non-monetary goals are often worse than their greed.
Ouch.. Carpet bombing with truth, not sure if legal.
It is refreshing to read a piece with a broad perspective that solely focuses on real issues, instead of keeping the layman in the dark with both-sidisms or "let us poke your emotional brain with political distractions". Puts corporate media like the NYT to shame; they could have saved millions on ho-hum opinion pieces if they wanted to.
Share this piece with your network, if you wonder what role you can play. Bottom up is the only way, the top is either clobbered with private interests or still having no other thought then "I believed we could be a bit corrupt, just pushing a bit but keeping it safe, no?"
> Honestly don't know what people think they are gaining with a heavy frontend.
True, but (I repeat myself here), it depends on what kind of website we are talking about. For instance, a data-heavy SPA that workers use the whole day (like a CRM) is at least perceptually faster and more user friendly compared to the same thing but with traditional whole page reloads.
I am open for suggestions, but anything wanting to give a desktop like experience is going to be complex. Like the user clicks a button, now widget a1 » a1.3 » a1.3.2 » a1.3.2.2 should be in an "open state", while widget b1 » b1.2 » b1.2.1 needs to be in "disabled state" and widget c3 » c4 » c5 shows a status message.
I am the first to admit that SSR worked fine since PHP. But two points I like to make
1) choose boring technology; 2) it is the library, stupid
First question should be: what type of web application are we talking about? If your blog does not render or your e-commerce site stops working without javascript, shame on you indeed.
If we are talking about SPA's, aka those replacing traditional desktop applications, and you want to throw out React, then we have something to discuss. Because, how are you going to replace MUI/MUI(x)? I keep hearing about the coolness of alternatives, and I am certain open to it, but in the end it needs to be not just better in some aspects or even fundamentals, while completely lacking in practical matters. A simple benchmark I keep going back to is to see if $hotness is able to replace [1], which shows you just one component of many in a cohesive system.
I don't contend with the principles in the article, but I do with conclusions like "ripping out react would be the solution".
checksums + tarballs don't help with runtime integrity verification. You'll need additional technologies for that like dm-verity or fs-verity; see composefs.
I am ignorant about this ecosystem, so I might be missing something obvious.
reply