somewhereoutth

Rather curious to know what is hiding in the 8.7B 'other expenses'.

Given that this is ~50% of revenue, it would have a huge bearing on how near real profitability Anthropic actually is.

Indeed and don't underestimate the influence you can have by participating in candidate selection (depending on your electoral system). For example in the UK the replacement candidate for a safe seat [i.e. almost guaranteed to be won by the party of the sitting representative] upon retirement of the incumbent may be decided by just a room full of people. Be in that room.

With a top heavy population pyramid, more resources have to be diverted to caring for the elderly, therefore less for raising the young - a self reinforcing mechanism.

When actually given the choice, many women are choosing not to have children, for good reasons. E.g. disruption of careers, physical effects on the body, being tied to a partner, financial drain.

Until we face up to both these issues, demographic collapse will continue - humanity will simply evaporate.

> A financial manager built a purchase request review agent

> An executive assistant, who built a personal assistant agent for scheduling and email triage

Really? We should build our own agents for everything? Sounds a bit token-spendy.

Also, as far as I can make out, the paper was not written by AI (nor even with AI assistance?). You'd think they'd make a song and dance about it if it was. Maybe AI can't do this sort of thing yet (or ever).

Wouldn't it be ironic if, after all this, it was discovered that the most economical way of producing some fundamental 'unit of intelligence' was to feed/clothe/entertain a human?

I like the phrase Work Shaped Objects to describe the output of gen AI (I thank The Tech Report YouTube channel for bringing that to me).

So - prose, code, or image, it appears that some work has been done, but in fact the [actually needed] work has likely not been done.

It sad that it has come to this, but your approach is correct - it is the act of writing (code or prose) that is important, not the finished output itself.

Interestingly, this applies not just for learning, but also to a great extent in production/work contexts too.

I went to buy a new phone about a year ago - all the branding/blurbs etc was "AI this! AI that!"

I didn't end up buying at that point, but 6 months ago my existing phone failed, so I had to return to the shops. Same phones for sale more or less, but not a single word about AI anywhere.

Also, writing, indeed any making activity, forces you to think clearly. Often that is the deepest purpose of prose, even code - forcing the author to confront the problem and sharpen their own understanding of the world we share. That it then communicates something to another person (or to a machine), is a happy bonus.

An awful lot of effort has gone into making LLMs present as human/intelligent. Without that they would just be a prose / code / image generator and/or search assistant, and nobody would be pouring 100s of billions into the technology with the end goal of replacing expensive human labour.

everyone (on all sides!) is losing at AI and we'll see increasingly frantic attempts from all players to paper over the cracks as this extreme resource misallocation bites.

and I suppose the rest of us will have to wait until the grim laws of finance do their work before our own sweet release.

Exactly this. The overall productivity of not doing something should never be underestimated.

This chart of free cash flow for the big tech companies should worry people:

https://www.advisorpedia.com/media/media/2026/02/23/trailing...

Somehow we've managed to turn software from a machine for printing money to a machine for setting it on fire.

However, who will pay for the very expensive training if the ROI isn't there? Absent government subsidization, then likely the frontier models will be sold off to the hyperscalers by the bankruptcy administrators to wring a few more dollars out of them as they become out of date and more and more irrelevant - assuming that inference alone is profitable at a price the market is willing to pay of course, if not then yes the lights will indeed be turned off.

Banning smoking from indoor spaces did in fact make it go away [from indoor spaces].

Actually, the analogy is compelling - an addictive habit that is pleasurable to the user, profitable for the supplier, but others around them must suffer the effluent with consequent deleterious effects on [mental] health. And similarly, the worst [mental] health effect is of course on the user themselves.

As with smoking, probably we can't/shouldn't stop individuals from indulging in this habit, but surely we can protect the rest of society.

The key point I'm trying to make is that the shade structure or outer shell should be separated by at least 50cm, and in such a way that even the slightest breeze moves air though it.

Great for low density single storey dwellings, not so helpful for medium/high density multi-floored residential.

But there is no airflow inbetween the outside and the building core - the panels are still part of the building itself.

Maybe we need shade structures - a shade above/around the building and detached so that air flows underneath and there are no thermal bridges to the building proper. Like you would make to protect tents at e.g. Burning Man.

Of course the tragedy is that the actual food had the 'AI' treatment years ago (in the US), the artwork is only now catching up.

I suspect it is a fair representation of the food you would receive - looks kinda like real food, but actually isn't.

the key comparison is variable reward.

That 'triggers a surge of dopamine and creates highly addictive habits' [thanks Gemini!]

LLM use for code generation does exactly that, sometimes it works amazingly, sometimes it fails inexplicably. Whether it is negative sum or not doesn't really matter. Indeed it may well prove to be negative sum, especially if we step back a bit and consider the business benefit of the code produced, not just lines of code or even features produced.

> and a few new tricks I discovered

This right here. Any gambler would recognize that statement.

So... basically watching them set fire to money in realtime??

The EU is predicated on the pooled sovereignty of its constituent countries, as exercised thought the EU Council. Apart from a limited number of certain matters, any EU country can veto any decision made by the EU Commission, or indeed the EU Parliament.

I went to look for a new phone about 6 months ago - everything was AI this, AI that, all the display stands made a big deal about AI (to the extent I considered a 2024 phone because no AI). In the end I didn't buy at that point.

Just last week finally I had to buy a new phone, and so I went shopping again - this time no AI was mentioned anywhere. Not online, not on the stands in the shops, nowhere. The silence speaks volumes, as they say.

*to billionaires

but snark aside, society needs to have a big conversation (meaning political) about what is good and what is bad about what should really be understood as the 'connectivity revolution' of the last 10-20 years.

was this before or after Lisp's epiphany for lexical binding?

Following S.I. notation, we might have Tk for token, so kTk, MTk, GTk etc.