Anthropic is running a similar marketing campaign as AWS/Devops tools which were trying to replace in-house IT. Pitch to the few that you can be 10/100x as productive and valuable on the hopes that they will push their organizations in this direction.
I dunno, but if someone is saying they expect to spend a vast, unprecedented sum of money acquiring an interdependent set of resources that current production could not come anywhere close to accommodating in multiple dimensions, and which money they don’t have and aren’t making in their current operations so that it would also require unprecedented fundraising on top of the other issues, you’d probably want to do some work to verify the plausibility before putting your own resources at risk for the chance of profiting from that spending spree.
Tesla and similar companies really make me wonder if we still live in a capitalist system. If wealth is sufficiently concentrated - the value of anything becomes tied to the whims of the few who can transact at that level.
How a stock goes up while sales growth, profitability, and other measures go down on a multi-year trajectory defies my understanding.
Same as any other economic system: power is usually concentrated around a very small group of people. In socialism and communism, that concentration typically occurs within the party leadership or central planning apparatus.
However, in free-market capitalism, anyone is allowed to participate in capital formation and accumulation. Ownership is not formally restricted to a political class. Entry into markets is open in principle (unless it stops being a free market), and capital allocation is decentralized through free and voluntary exchange rather than administrative decree.
That does not mean capitalism eliminates power concentration, as Wealth can accumulate and translate into political influence. But the mechanism of power differs: In centrally planned systems, control flows from political authority. In market systems, control flows from voluntary transactions and competitive success.
> But the mechanism of power differs: In centrally planned systems, control flows from political authority. In market systems, control flows from voluntary transactions and competitive success.
This is a boilerplate theoretical explanation of capitalism which bears very little resemblance to how things actually work in most Western capitalist societies, particularly America.
A cursory glance at the majority of large markets reveals that they are dominated by a handful of massive companies, who use their enormous wealth and influence to steer (centrally planned) political policies, thereby stymying competition and entrenching their position.
I don't follow. Even now there is nothing preventing anyone here from making something for millions of dollars. While VC capital is closed to a select few, a person in a garage can still make it big.
Communist counties tend to gate keep even more. To the point that it is entirely who you know, with little concern to what you do.
Only if you limit yourself "capitalism" and "communism" as the two economic systems you are are considering. What we should be doing is noticing that these two systems fail in very similar ways (concentration of power in a small group of people), and think about what kind of system might not fail in that way.
One of the things I’ve noticed is that when leftists say “capitalism” they often mean “the ability of capital to set the rules of markets” rather than just “markets”. This causes people who use the latter interpretation and leftists to talk past one another quite a bit. Which is one of the reasons that leftists have sounded this alarm bell for at least twenty years and no-one has paid attention.
Capitalism is a big money party. Leftists are the party poopers, actually just slightly less drunk than the rest of the guests, and pointing out that lighting up fireworks indoors isn't a good idea. Booo-hooo, shut up lefty! *BANG*
Command economies break too. Capitalism has occasional fires like a forest. This is good. It allows things to shake out. New growth comes from the destruction. Command economies don't. They just continue to grow weeds and tall trees until everything is choked. Then they wonder why everything is garbage.
The US is moving to a fascist economy. That is a form of command economy. For example the FDA is controlled by big pharma.
Also - how can this be prevented? the AI labs can't seriously expect that each lab will filter LLM generated content from their training sets based on the source model. Leakage of AI behavior into public datasets is inevitable.
The question is how much further and faster can these tools evolve.
Right now, you may not be faced with a valid choice to vibe code your own hubspot, but maybe some contractor firm will do it for you and sell an ultra low cost version.
Will opus 6 do the same thing without the contractors while managing the deployment and maintenance as a claw?
Curious if part of this was the overall decline in government compensation relative to the private sector. The president makes roughly what the typical SV engineer makes after 5 years in big tech or as a fresh grad from a top PhD program. Meanwhile the people the president deals with have become unfathomably wealthy.
In 1909, the US president made 75k - roughly 2.76 Million in today's dollars. This is in comparison to the current 400k dollar salary of the president. As the president is the highest paid government employee by law/custom - this applies downward pressure on the rest of the governments payroll.
I see no reason why the president shouldn't be modestly wealthy given the requirements or the role and the skill required to do it well. Cutting the payscale to less than some new grads seems like a recipe for corruption.
Since 1958 with the Former Presidents Act [1] the Presidency guarantees you'll live very comfortably for the rest of your life with a lifetime pension (and even a small pension for your wife), funding for an office/staff, lifetime secret service protection, funded travel, and more. It was passed precisely because of the scenario you describe playing out with Truman who was rather broke, and ran into financial difficulties after leaving office.
Your source does not seem aware of the history of tax brackets. In Truman's era the top rate started at $100k income and was 90%. Even at $60k it was 80%. Those are figures from 1953, when Truman left office. [1]
This wasn't that big of a deal for the average person at a time when the median salary was somewhere around $3k, but for a person with significant overhead and large, but not enormous, income -- in other words, exactly like Truman -- it was devastating.
> Your source does not seem aware of the history of tax brackets.
"Beginning in 1949, the president was also granted a $50,000 (equivalent to $677,000 in 2025) expense allowance, which was initially tax-free, and did not have to be accounted for."[0] That's 4 years of $677,000 tax-free which I make to be about $2.7M which lines up with "Truman embezzled about two and half million dollars, in 2025 money, from the White House expense account"[1] - "tax-free", "did not have to be accounted for" -> tax brackets are meaningless.
But, you say, "the allowance became taxable later in his presidency"[0], and I reply "Truman never reported it on his tax return"[0], "and also didn’t pay the taxes he owned on the money."[1] which also somewhat scuppers the "history of tax brackets" angle, no?
Further, "In February 1953, Truman signed a book deal for his memoirs, and in a draft will dated December of that year listed land worth $250,000 (equivalent to $3,008,000 in 2025), savings bonds of the same amount, and cash of $150,000 (equivalent to $1,805,000 in 2025)"[0] which I reckon comes to about $8M which, again, lines up with "Truman had a net worth of about $8 million in 2025 dollars when he left the White House"[1].
Further, further, "In January 1959, Truman calculated his net worth as $1,046,788.86 (equivalent to $11,561,000 in 2025)"[0] which, to be fair, is slightly lower than the "$14 million in 2025 dollars when he was successfully shaking his tin cup to Sam Rayburn and John McCormack in 1958"[1] but in the same "NOT AT ALL POOR GTFO" ballpark.
In summary - he embezzled $2.5M, got $8M for your memoirs, and ending up being worth $11-14M within 6 years of leaving the White House and thus I rate the claim "Truman who was rather broke, and ran into financial difficulties after leaving office" as 100 Pinocchios.
Wikipedia is not a source. All of the dirt on that page is exclusively relying on one article [1] from a self-citing author, the same author you again cited in your post here, from 'lawyersgunmoney.' I don't find it compelling. It may indeed ultimately turn out to be accurate, but I find it generally unwise to rewrite history on the words of a single person, who seems to have a strong bias towards a certain narrative.
That bias could indeed just be because he believes he's discovered a truth which most people don't know, completely contrary to the 'official narrative', which is indeed quite a frustrating place to find oneself. But it can also cause one to be blind in some ways. For instance assuming everything as written is accurate, the author simply then jumps to malice (like tax evasion), seemingly without consideration of issues such as inconsistent or flawed record keeping which I imagine was extremely common in the days prior to computers.
Are most fresh grads from a top PhD program really making $400k/year? Sure, the ones hired by OpenAI are making at least that much, but the vast majority are not. However the broader point remains, that the president’s (and the rest of government’s) pay structure has not kept up with the private sector.
It's quite plausible to me that the difference is inference configuration. This could be done through configurable depth, Moe experts, layers etc. Even beam decoding changes can make substantial performance changes.
Train one large model, then down configure it for different pricing tiers.
I dont think thats plausible because they also just launched a high-speed variant which presumably has the inference optimization and smaller batching and costs about 10x
also, if you have inference optimizations why not apply them to all models?
This really points to a world where all services are too cheap to meter. The compute side of AI is a commodity, the usage of AI is a commodity, the model development of AI is a commodity. So far there is no evidence that a provider with heavy usage has any long-term advantage over a vendor with no usage. New top tier models come out every week from relative unknowns.
Other than a vast consolidation of what parts of the economy are "digital", what is going to have margin other than orphaned capital and "creative" efforts within 10 years?
EDIT: the top ranked model on openrouter based on traffic changes almost weekly now, I can't see how Amy claim of “stickiness” exists in this space.
It's a good way for a grocer to minimize waste. When raw chicken gets close to sell by date, turn it into rotisserie chicken, when it doesn't sell - turn it into sandwiches and other products.
reply