I wouldn’t care about the rate limit if it didn’t make so many mistakes. It’s in their best interest to make mistakes too so you use more tokens

0% my dude. There is no lvl 5 driving car. Anything below lvl 4 requires human supervision (how often is a different topic).

Source? Has it progressed so much? I remember substantial double digit percentages.

It’s just managers who have long since abandoned retaining any practical knowledge in the field, if they even had any at all, and pushing what their boss wants to hear and what they have all convinced themselves to be true in the world of management and linkedin lunatics, because that’s how you get promoted in management. Because everyone from the top down is all in on AI because those at a high enough level are all chasing a seat at the table of the ownership class holding the keys to the AI kingdom and the stock valuations that come from it.

So if their company or their company’s AI model or AI usage makes them an important client of someone that the stock market has deemed important, or if they think their company is the special chosen one that will figure out the secret to AI productivity that none of the other thousands of companies will, that is all that matters. They know nothing about its actual usage, just that more tokens = more usage and therefore must be good, because more usage means their company stock goes up and it that means a bigger golden parachute.

The problem is that either they are genuinely ignorant of the fact that it is a bubble, or they believe that they will be able to time their exit right. The worst part is that a lot of them will, but all the people and the organizational chaos that the burst of the bubble will create that they are knowingly contributing to, well that’s some other sucker’s problem. And that someone else is everyone, whose power and water bills will go up, whose ram and hard drive prices go up, and whose 401k will tank when the bubble collapses.

In my experience with Claude Code Opus, analyzing or generating files eats the most tokens while thinking is actually surprisingly cheap. I guess, the token-counting is somewhat wrong on purpose to incentivize use of high-effort thinking mode because when you incentivize using lesser models or modes, people get disappointed by the output quality and stop using the service…
So just let it analyze the code base for flaws and bugs in a loop using lots of sub agents for each type of bug or code smell.

The good thing about that method is that it is technically malicious compliance; but it also offers a high degree of plausible deniability and likely yields some actual bug fixes to offer as justification.

Doing it just once every once in a while without fanning out into tons of agents rereading the same files is the non-malicious-compliance way of using AI for bug hunting and usually worth it. Also let it write tests for the found bugs (and properly review those tests using the natural neural network in your head).

Probably just ask it for the seahorse emoji or something idk

Only if it’s been trained on discussions about it.

It’s kinda like the push to return to office. It was driven by corps having invested in the “can’t fail (ignoring the last previous crash)” real estate market and buying their offices. If everyone suddenly works from home instead of in the office, then those investments go bad because demand for office space is way down. So they tell people to go back to the office, hoping to return to that “every business needs offices!” status quo and save their investments. Though the demand is false (especially combined with layoffs), so it won’t necessarily cause any new corp to want that office space. If they don’t have the sunk cost, then they don’t need to accept the rest of the fallacy.

With AI, it’s the same but just replace building investment with R&D as well as data centre investment. A lot of the companies really pushing AI are the ones that will profit from people going along with that. They really want to build a dependence amongst users as well as a good reputation for execs so they can get a return on the investment. Then there’s also the True Believers (who think LLMs are brilliant AIs that can solve anything if given the right prompts) and the FOMOs (who don’t know much about it but see the world moving towards it and don’t want to miss it because if it was a real AI, missing it could be a massive mistake). There’s also some people who just don’t have various skills and want the AI agents to fill those gaps (and probably don’t have a very good idea about what the LLMs are actually doing in those gaps).

At this point, I think it’s a mistake to go all in on this tech. LLMs aren’t reliable, and their ability to “perform” is more about their flexibility than being well-suited for any task. They’ll go directly from saying things that seem “insightful” (they have no insight) to making the dumbest “mistake” (a mistake requires intent, which they lack, they just predict tokens). But there’s all kinds of false and true (albeit misguided IMO) demand right now and it’s still in early pricing mode (remember the intent is to make that investment money back).

Oh and there’s also China which has been making more efficient models and open sourcing some of them. If they continue to do this, there’s a decent chance those investments will never give the desired returns, at least not to those who are trying to sell tokens. Or those who depend on those selling tokens, like any hardware companies selling hardware under the assumption that it will then make the money to pay for itself (which I believe both nVidia and AMD have done).

It’s mirroring the dotcom bubble with that last bit because network cable companies started loaning the money to pay for their cables to ISPs, expecting returns that never came.

I don’t favor forced adoption… but your second sentence just isn’t true of individual people. It is ture of the industry in general. But individual people tend to stick with what they know and shy away from new things until something forces them out of that state. Usually it is someone else who tried a new tool and is suddenly able to do things easier or what not. It only takes a few power users to turn the tide. So forcing shouldn’t be needed. Just enabling those power users.

Huh, it’s been an operating expense since, let me check, the dawn of business?

“Investing in” your employees doesn’t make a ton of sense when they could just hop over to the next company after a year, much like loyalty in a company doesn’t make sense. It’s all a business transaction in the end so if you can get a better deal, you should.

The unfulfilled promise of AI is to reduce the expense to a minimum so the few remaining people can accomplish the same work. That’s more of an investment than paying people to work for you until they find a better job ever was. In theory.

Uh they’ll just claim the AI is being used in another country then?

There’s really no way to definitively decide the tax jurisdiction of an AI agent.

Good luck figuring this one out without a tax on unrealized gains in companies and good luck figuring that one out for companies if they can just scheme around it by reinvesting all their profit indefinitely.

A special 50% AI services tax for the company selling access to the AI might work though. Like VAT, but don’t let companies claim it back. Good luck defining what AI is legally though.

You can’t increase taxes on companies. They’ll leave the country.

Total cost of employment is something like 140-160% of salary for a software dev. So my super handwavy half pulled out of my ass math puts it closer to 20 jr. engineers.

You forgot the 1/3 part (hope that is the brain part)

Less than half of what I hoped for.

Compression doesn’t add/change hair.