edschofield

Accounting 101: CapEx doesn’t directly affect either gross or net margin (when the expenditure occurs). It is capitalized on the balance sheet and affects profitability over time (through depreciation or amortization).

[1] https://en.wikipedia.org/wiki/Capital_expenditure

$100b of annualized profits at a 50x P/E ratio?

I think that's possible if Anthropic manages to ban open-weight models in the US. If they can't manage that, I don't yet see how they will reach profitability ...

The same Google that drowns you in low-quality AI responses, ads, and “organic” traffic stuffed with ads? I’d prefer the private, local, homegrown specialized English-to-bash translator…

Not really. Kagi has its own indexes and says a typical query consults about a dozen sources. [1]

[1] https://help.kagi.com/kagi/search-details/search-sources.htm...

I had a great time at PyData Amsterdam last year. Terrific people!

I thought this post would be about something that to me is so significant and obvious but I have never seen discussed: that labs like Anthropic are deliberately misaligning their models with their users’ goals. Fable’s refusal to secure your codebase is HAL 9000’s “I’m sorry, Dave. I’m afraid I can’t do that.” Whenever you read “alignment”, the question is “alignment with whom”?

The parent commenter described exactly how Macs work. I have a beautiful 10-core iMac 27” with 72 GB of RAM from 2020 that Apple has now decided to turn into a lemon. macOS Tahoe runs as slow as molasses and no more macOS upgrades will be forthcoming. I downgraded to Sonoma (which cost me hours) and the machine runs well again for now but the App Store has already started refusing to update apps like Pages because macOS is old. Soon I won’t be able to install a single app through the App Store and Linux (T2 Linux in my case, not Asahi) will be the only option to keep using this terrific hardware.

I had the same experience as you. I bought a super-fast 10-core iMac 27” with 72 GB of RAM in 2020. It ran beautifully until the moment I “upgraded” to Tahoe, when it instantly became unusably slow.

This cost me half a day wiping the disk, restoring the original OS, and then carefully upgrading back to Sequoia. Now it works well again, but I am counting the days to when Linux is the only way to continue using this great piece of hardware.

Another anecdote: a few years ago I was offered a shiny new never-before-used iPad for free by a client (a telco). There was only one catch, I was told: it was a few generations old. It had sat in its box on a shelf unopened for several years. I thought it might be useful at least as an e-reader, but I soon realized why it was now useless: I could not find a single app in the App Store that would install on it because every app required a “newer OS” (which wasn’t available). Most websites also refused to talk to its old (non-upgradable) web browser. Well played, Apple.

Thanks for the laugh!

Claude Code does now have integration with Telegram, Discord, and iMessage as of a few weeks ago:

https://code.claude.com/docs/en/channels

I haven't used OpenClaw since then ...

Any company worth more could (in principle) acquire it with a share swap [1]. Even a smaller company could buy it with an LBO [2].

[1] https://en.wikipedia.org/wiki/Stock_swap

[2] https://en.wikipedia.org/wiki/Leveraged_buyout

This is not an edge case. Most HN commenters describe the latest two versions of macOS as being objectively worse than earlier versions: slower, less stable, more broken. There are significant numbers of “power users” who deliberately avoid upgrading or have actively downgraded macOS to Sonoma because they care about their computing experience.

I like “Clawboard”.

The design of Pandas is inferior in every way to Polars: API, memory use, speed, expressiveness. Pandas has been strictly worse since late 2023 and will never close the gap. Polars is multithreaded by default, written in a low-level language, has a powerful query engine, supports lazy, out-of memory execution, and isn’t constrained by any compatibility concerns with a warty, eager-only API and pre-Arrow data types that aren’t nullable.

It’s probably not worth incurring the pain of a compatibility-breaking Pandas upgrade. Switch to Polars instead for new projects and you won’t look back.

Yes, ChatGPT 5.2 Pro absolutely still does this. Just ask it for a pivot table using Polars and it will probably spit out code with Pandas arguments that doesn’t work.

Thanks for the link! I didn’t know Monday existed. I laughed so hard at its output. But I fear that using it regularly would poison my soul…

Numerical integration methods suffer from the “curse of dimensionality”: they require exponentially more points in higher dimensions. Monte Carlo integration methods have an error that is independent of dimension, so they scale much better.

See, for example, https://ww3.math.ucla.edu/camreport/cam98-19.pdf

Here's an interesting way to understand it:

https://pricedingold.com

The true value of gold is quite stable over time (since new gold is mined at a slow rate). Fiat currencies are constantly being debased. Hard assets fluctuate up and down relative to gold.

That's a Dutch auction: https://en.wikipedia.org/wiki/Dutch_auction.

I agree that reverse auctions would be a simple solution the current scalper problem. If demand outweighs supply, scalpers drive up the prices toward their economic equilibrium anyway, making ticket prices just as "unfair" to the poor, but with additional problems of trust...

“The Price of Peace: Money, Democracy, and the Life of John Maynard Keynes”, by Zachary D. Carter.

> the unpopular touch bar which initially removed a physical ESC key.

I still have the scars: I no longer use a Caps Lock key, even on my newer computers — it’s permanently mapped to ESC…