Hacker Newsnew | past | comments | ask | show | jobs | submit | tumdum_'s commentslogin


I wonder what was total token cost of the rewrite.


I think around $300-400k, but that does not include the github issue/pr/ci stuff that runs continuously, which could double or triple it. Somewhere, the first two weeks were said to be about $200k.


That total assumes API billing rates, which isn't what Anthropic pays for the inference. So I'm confused why people keep quoting that. It's not representative of what it'd cost for you either because you could use cheaper models.


> That total assumes API billing rates, which isn't what Anthropic pays for the inference. So I'm confused why people keep quoting that. It's not representative of what it'd cost for you either because you could use cheaper models.

It is, because this rewrite was marketing for Claude. So it reflects the price for what other companies might need to pay if they want to rewrite something else with similar size.


Yeah, when you spend billions in infrastructure you get to save $300k occasionally


From the Rewriting Bun in Rust blog post: 5.9B uncached input tokens, 690M output tokens, 72B cached reads, came out to ~$165k at API pricing.


That was the initial work, but since then there was lots of work performed with LLMs.


It tells you little about living conditions but it tells you a lot about importance of a country as a whole. GDP per capita of China might be far from US but due to its size it still makes China very important actor.


Or you can just use any of the actual European companies (I’m using Tuta).

https://european-alternatives.eu/category/email-providers


Any providers that don't provide SMTP/IMAP/JMAP are just a boring way of giving away control over your mail and the client software you use.

I don't know about Tuta specifically, but Protonmail is practically intentionally hostile against anyone using their own keys. Which is the biggest sign that their marketing and actual intent do not align.


I started using tuta until I realised they don't support IMAP. Something to do with not guaranteeing encryption (which isn't even enabled by default) but has the convenient effect of locking you into their apps


https://mailbox.org/ is great. It's the only provider recommended by privacyguides.org that supports IMAP/SMTP.


Tuta is always encrypted I don't know where you got the impression that it was optional or that they could somehow magically make it work over IMAP without a bridge like proton.


> Tuta does not support the use of third-party email clients or the protocols IMAP/POP3/SMTP as we cannot guarantee end-to-end encryption of your data.

So it "breaks end to end encryption" even with smtps and imaps apparently. The few emails I received weren't from tutamail users so presumably came over SMTP.

It's confusing to know what they mean because they confuse terms. They say emails are "stored end to end encrypted".

They don't pass my smell test


And here I thought using open source client software is the only way I, the user, can guarantee E2EE of my data.


They offer similar product to a more well known Proton.


PGP works over IMAP.


IMAP can do TLS though (IMAPS).

Did they say what's stopping them from using that (and requiring the encryption!)?


They didn't need to. When you start denying IAMP to your customer in the name of "encryption" at that point it becomes privacy theatre, instead of privacy, irrespective of how nobly activist their intensions are. It is probably slightly worse than a mail provider assuming they can't trust their users with encrypting their emails when needed.


> When you start denying IAMP to your customer in the name of "encryption" at that point it becomes privacy theatre, instead of privacy, irrespective of how nobly activist their intensions are.

How would the lack of IMAP compromise encryption? Sounds dubious.

Lock-in effects are problematic, but separate from encryption or privacy in general.

If the concern is actually lock-in, it is worth noting: there is little lock-in in the current Tuta offering. The service allows users to export emails into standard EML files. If you decide to switch providers, simply point your domain to the new provider and import your latest backup.


Its probably about end to end encryption. Protonmail doesn't store mailboxes in plaintext, only the client actually decrypts the mailbox. IMAPS only encrypts the conn between mail server and client.


Infomaniak is another one https://www.infomaniak.com/en/ksuite


It’s actually better to be average Western European than white and top 1-5% American: https://jamanetwork.com/journals/jamainternalmedicine/fullar...


It turns out that when you make a massive profit motive to sell expensive fancy interventions, health outcomes suffer.

You see this to an even more extreme degree amongst the ultra rich and powerful, who are often using cutting edge, experimental treatments, and often reaping the downsides of being on that edge (though of course also the upsides).


To be fair, the structure of that study is pretty bizarre. Why only look at 6 specific illnesses rather than something like life expectancy?


They are looking at the outcomes of specific procedures because they are interested in the health care system quality. Looking at things like life expectancy would make it impossible to disentangle them from other factors like lifestyle or environment.


Sure, but it's only six (6) specific illnesses. There are tens of thousands of illnesses. I'm almost 100% confident you could cherry pick any six illenses to get any outcome you want.

Edit: Seems the lead author is also a vocal supporter of universal healthcare, being a senior fellow of the Center for American Progress. So yeah, probably why it's such a strange setup, and what compelled me to check since this kind of thing is super common for "activist research".

I guess the takeaway is to always take a study for what it is, rather than what it paints itself to be.


So you are saying that they cherrypicked some obscure illnesses to make Europe look good? They used data for:

- infant mortality

- maternal mortality

- colon and breast cancer survival

- heart attack mortality

- childhood acute lymphocytic leukemia (most common childhood cancer)

Those are one of the most common health issues.

Also, here is another recent paper showing how it’s better (health care wise) to be average in EU than wealthy in US: https://www.nejm.org/doi/full/10.1056/NEJMsa2408259


No, I didn't make the argument they were obscure, in fact they all are from a set of common illnesses that....also contains thousands of illnesses. Again, this is the kind of stuff you see with "activist research". Why would they only do 6 instead of hundreds or thousands? I'm sure you see my point.

There are also other oddities like using median county wealth rather than household wealth. In many wealthy counties there are still poor areas, I know more than a few. In fact if you look at the appendix, most of the households in the study are, well, average or poor. They also don't include absolute numbers or definitions for the illnesses.

All in all, it's a bad study, and clearly made to color in an agenda.

The second study looks more grounded, thanks.


> The reallocations were real, but they were never the bottleneck.

Why do people not write their own blogposts on their own?!


Latest presentation of Terence Tao on what current advancements in AI mean for math discusses (among other things) those issues: https://teorth.github.io/tao-web/slides/age-of-ai-icm-2026.p...


Thanks for sharing. Regrettably his conclusion on proof exposition increasing in importance reflects the broad trend in every field: When someone (or something) else is doing the actual work, all that's left for the original folks is to find a way to sell it. Pretty soon mathematics will be awash in marketing with slogans such as "I am a famous mathematician, I checked this proof and it looks legit. Trust me." Absolutely nothing wrong with that scenario right? Except for the logically fallacious appeal to authority and the inevitable corruption over time.


No. More like "I am mathematician and I have digested this sloppy writing and understand this to be an application of Foo theory to the Bar problem with a Baz twist. The prior publication omits crucial references to..."


The writing's already pretty good, even Terrence Tao calls it flawless, if you read the slides. That's not where the LLM requires help.


Is google still a big contributor to the c++ standard? I thought they are far less involved than a decade ago.


I'm not sure how true is this claim about weak environmental laws. I don't have any data to back this up, but I was in China in 2013 and two years ago and the difference in air quality in major cities is staggering. So, maybe they destroy the environment but for sure they no longer do that near the big cities.


Reading a book is a way to peek into author's mind. There is nothing to peek into when the book is generated by llm.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: