throwawayffffas

Hates users who are not total normies and designed explicitly for normies are not opposite. In fact an app that hates not normies would be designed for explicitly for normies.

> Easily? There's no easy in maintaining and releasing two different versions of your app, especially for a small non-profit.

It's literally just another job on your CI. They got a 50 million seed donation in 2017 and reported almost 30 million in revenue in 2024. They are not exactly small. If they were VC founded and not a non profit they would be a unicorn.

I remember SO answers almost always looked superficially correct but were almost always wrong in a nuanced way that required experience to divine. Kind of like circa 2025 chatbots.

But at least we had fun who can forget stacksort.

Because they are swamped and want to automate some of the process.

The way I see it the devs that put that in place are saying, listen before even reading your contribution I am going to feed it to a LLM with these instructions and see what it spits out. Now I will extricate my self from being the LLM proxy so you deal with this output and when it's okay I will look at it.

Same as "listen bud until it passes linting I won't even look at it." Just a much more flexible and fuzzy linter.

Eh, you read compiler and linter output don't you? Nobody read that before you did.

Sure LLMs are not nearly as rigorous about their output but that doesn't mean it's entirely not useful.

At a minimum you can have your LLM summarize the other LLMs output and go from there.

I don't know what you have to do in terms of letting the grid know. But local authorities have stringent regulations about installation and almost definitely require permits and installation or inspection from vetted electricians.

The question is not around classification but of ownership and especially of the ownership of the documents after the president leaves office. The argument is that the documents generated during the presidency are the personal property of the president.

My hypothesis is not that the AI is going to eat up that 12T dollars. But there will be additional investment to provide AI tools to the workers to the tune of 10% to 20% of their salaries.

I think the market has the potential to be 1 or 2 trillion in the US, 10% to 20% of labor costs. I say this because I believe most companies would not blink at spending that money to supply their workforce with AI tools assuming the AI actually did the work.

The winner take all network effects are probably not there in this case, but still if Anthropic is a large player and has 20-30% of the market they could have a revenue in the hundreds of billions.

Double the size of the market to include most of the rest of the world and the 500B is suddenly under 20% market share.

The money is not on the casual users, the money is on the lawyers and accountants, and consulting specialists that have actual business use cases.

If you are spending 50 - 60 grand a year on a paralegal will you balk at another 5 or 10 that will allow them to do double the work?

I think you are underestimating the market for AI. The better it gets the more the market expands.

The total amount spent on salaries in he US alone is over 12T dollars.

Assuming AI becomes ever more useful and is widely adopted 500B in revenue is not that crazy.

The way I read it, it does. It looks like total revenue vs total compute expenditure.

Just renting them out as hardware is close to 12 months.

Given the interest rates a valid explanation is that we are not done with the employment contraction still.

But it looks to me that the layoffs are due to AI but not because the positions are replaced, because capital is flowing to hardware instead of new hires. The growth engine of the tech industry used to be people, now it is clearly gpus. If you have 10 billion in the bank you are not going to aggressively expand your workforce, you are going to build datacenters.