Berkshire is sitting on nearly $350 billion in cash. One of the ratio's that Buffet uses to measure overvaluation, Wilshire 2000 Market Cap as a % of GDP is at an all time, 220%. Its usually a good indicator of a market top. He's making a few small investments, but he's just waiting until there's more of sell off before heading back in. He's under no pressure to buy anything.

It'll come from the other bubble in financial markets, Private Credit. There have been quite a few bankruptcies and problems emerging there, and they're stepping up to be major funders for the AI capex boom.

APPL is the best bet as it doesnt have any exposure to AI. MSFT has spent a truckload of money on datacentres, if AI doesnt deliver they're going to face a huge drag on earnings because of the overinvestment.

Microsoft has the same struggle as Facebook and Google.

If AI matures into the next big thing, it risks wiping them out of their dominant positions if they don't spend to keep pace.

If it doesn't, then they just blew a bunch of money but are still dominant players, with an existential threat in the rearview.

Given the relative risks, they'd be foolish not to pump some of their cash flow into AI for a few years until the pace settles.

Meta and Alphabet are mostly ads + desperately trying to find other ways to make as much money as ads

If there is a bubble then the hyperscalers with dominant market shares and big cashflows are most likely to survive if it bursts. Google, MSFT and Amazon. Oracle are on very shaky ground. Their bonds are selling at a discount to par, even though interest rates are dropping, and their debt to equity ratio is nearing 400%. AI is an existential bet for them, but not so much for the others.

First it was Softbank and now its Thiel. Major sales coming from two of the ultimate tech insiders is probably a very good sell signal. Whether or not its the top is another question. Markets as we know can go higher longer than we think, Michael Burry shut down his hedge fund because he may have started shorting too soon.

One of the funds that Meta were working with to finance their spending on AI, Blue Owl, had to stop redemptions on of their private credit funds merging it with another one of their bigger funds, leaving holders facing a 20% haircut. There also seem to be emerging liquidity issues with banks. The money for the AI gravy train might be running out.

didn't softbank sell only to invest more into other ai investments?

Yes, but whether they will remains to be seen. They needed a reason to sell out without it looking like they were selling out at the top.

I thought I read specifically with the intention of putting it into OpenAI directly, but yes they said they wanted to “pursue other ventures in the area of AI”.

If you're a finance professional then there's no alternative to Excel. But otherwise Libre office calc is perfectly fine.

> If you're a finance professional then there's no alternative to Excel

Not sure what you mean by this exactly, but I work in banking with a lot of "financial professionals", and the general opinion is that Excel is not good because it screws with numbers, whether its scientific notation (Why? Its just as long as the original number), rounding of numbers (had that with a large list of account numbers just last week where half the account numbers lost the last 3 digits) and there is no easy way of saying "just treat these as entered".

Even setting fields to text doesn't stop Excel from fucking around and overriding them to be date formatted if it feels like the balance could be.

The main issue is that Excel comes with Office and you aren't allowed to install other software so it forces you to use it and get used to it. It really wouldn't take much to be better than Excel.

The problem, as I see it, isn't a feature problem but rather just the fact that everyone in banking and finance is already using excel. You aren't going to see `ods` files passed around.

I would agree with this.

If your bank is exporting and importing csvs, the problem is on the software stack side.

Both xlsx can be exported and imported. It is just harder

The only reason this is true is because everyone in finance uses Excel which means that differences in parsing excel docs is consequential. And, in finance, excel docs get shared a lot.

It's not the case that calc is lacking any features which excel has in a finance situation.

windows - if you need kernel anticheats, excel or CAD

mac - if you need battery

Sometimes it feels like Google are so far ahead in AI but all we get to see are mediocre LLMs from Open AI. Like they're not sharing the really good stuff with everyone.

I think I believe OpenAI's claim that they have better models that are too expensive to serve up to people.

I think Google have only trained what they feel they need, and not a megamodel, but I can't justify this view other as some kind of general feeling. They obviously know enough to make excellent models though, so I doubt they're behind in any meaningful sense.

This has been said by enough people in the know to be considered true by now. Not just from oAI, but also Anthropic and Meta have said this before. You train the best of the best, and then use it to distill/curate/inform the next training run, on something that makes sense to serve at scale. That's how you get from GPT4 / o3 prices (80$/60$ /Mtok) to gpt5 prices (10$ /Mtok) to gpt5-mini (2$ /Mtok).

Then you use a combination of the best models to amplify your training set, and enhance it for the next iteration. And then repeat the process at gen n+1.

When they do, someone like OpenAI comes along and runs the VC enshittification playbook on it.

Remember "Do no evil"?

This sounds like another bogus role they'll ditch once they get their Nasdaq listing and need to make profits for their shareholders.

I'd probably trust any organisation with a role like this even less. It sounds like an organisation that doesnt think it can be trusted.

I remember "Don't be evil"

How'd you get so cynical?

Wisdom acquired by age.

Youve never seen enshitification happening? These companies butcher their customers as soon as their incentives shift to increasing short term shareholder value

They're trying to wreck as much of the current governmental set us as they can do it'll almost impossible or very difficult to rebuild it. It's almost scorched earth, they think they're killing the "deep state"

I think the "deep state" crusade assumes a sort of good faith that it's obviously lacking in this administration, judging their intent from their behavior and outcomes paints a much scarier picture.

You only need “good faith” from voters and supporters, the people in power know exactly what they are doing.

We created this oligopoly because they were convenient, free, powerful, and now its time for us to pay the price.

Or find services that may not be as easy to use, may cost something and may not have all the features you want, but which wont make unreasonable demands for your data.

In light of the way the US government is carrying on, I'd rather not give Microsoft any of my images.

> In light of the way the US government is carrying on, I'd rather not give Microsoft any of my images.

What is this supposed to mean? That you'd be happier with the dystopia if they were going after people you like less?