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

This was not what I was expecting. The doctors I know are mostly miserable; stuck between the independence but also the burden of running their own practice, or or else working for a giant health system and having no control over their own days. You can see how an LLM might be preferable, especially when managing a chronic, degenerative condition. I have a family member with stage 3 kidney disease who sees a nephrologist, and there's nothing you can actually do. No one in their right mind would recommend a kidney transplant, let alone dialysis for someone with moderately impaired kidneys. All you can do is treat the symptoms as they come up and monitor for significant drops in function.

Perhaps we didn't realize how much stability the "two powers" model generated. It caused inevitable arms races as the two powers vied to stay competitive, but there were only two. And the USSR was able to de-escalate on its own. If you have three powers, each of them wants the ability to eliminate not one, but both of the others. Could lead to not just incremental, but polynomial expansion of forces. And de-escalation involves multiple parties coordinating, not just one great power.

If you have a weird phoneme / meaning mapping brain like mine, I would note that he is not the doctor who is known for the "replication crisis". Even though Ioannis means John in Greek. Took me a second to tease that out.

https://en.wikipedia.org/wiki/John_Ioannidis


It does make you wonder what would cause a downgrade. The debates over the debt ceiling have certainly brought the U.S. closer to default than I would ever have thought. It's true that the U.S. can never run out of dollars, so in once sense it's not possible for a bondholder not to get paid back. But the political environment, the potential unreliability of previously iron-clad data, economic disruption from tariffs, and behavior from the Federal Reserve, these all seem to make an unlikely event much more likely.


Well, they wouldn't be able to pretend that they are selling from the official store for that inventory. Which I, personally, would be OK with. I've been on eBay for a couple decades, I don't mind ordering from Jack and Jill's Computer Parts as long as they have a reputation I can check. But the current situation where you can order from what looks like the the official storefront but the fulfillment is from a seething mass of "stickerless commingled inventory", with no way to even determine which merchant introduced the counterfeit product? This has been a problem for over 10 years. It's not just the obvious fraud, it's the subtler fakes. I won't buy anything from Amazon where the failure could kill or injure someone. A sun hat? Sure. A charger or food? Not a chance.


> I won't buy anything from Amazon where the failure could kill or injure someone.

you can do better, don't buy anything from amazon at all, ever

works for me


Oh, I have a long list of vendors that I'll buy from over Amazon. I buy almost nothing from them. On the rare occasion that I simply can't find something locally or from a reputable vendor, or need it on very short time scale, well, OK. But we dropped Prime, where we were ordering 100+ times a year, and now I pay out of pocket for shipping on a half-dozen orders a year.


It's not clear what LLMs are good at, and there's great interest in finding out. This is made harder by the frenetic pace of development (GPT 2 came out in 2019). Not surprising at all that there's research into how LLMs fail and why.

Even for someone who kinda understands how the models are trained, it's surprising to me that they struggle when the symbols change. One thing computers are traditionally very good at is symbolic logic. Graph bijection. Stuff like that. So it's worrisome when they fail at it. Even in this research model which is much, much smaller than current or even older models.


It was the same thing for us with Qt Commercial licensing. We use only the LGPL version, dynamically link, don't modify the source, and give credit, so we're fully in compliance. To get support we chose to purchase commercial licenses for our small team of developers. Cue a regular series of calls about whether we were sure we were in compliance, etc. To add insult to injury they couldn't even navigate our purchasing process so it was a pain to pay them.

I'll take my chances in the open source world. It's a shame that the companies that created the software aren't getting paid, truly. But don't make it so obnoxious to reward you.


I chose Migadu because they seem to be genuinely helpful and are very affordable. I probably would have gone Proton but they don't support forwarding.

The downside is that downloading messages is fairly slow when you have 10-20k messages in your inbox. And the webmail is fairly primitive.

I never tried Fastmail.


I’m also a happy Migadu customer, and I have never had to interact with their support. I use it for most of my domains, except for a few that are work-related.

For Webmail, I have been meaning to try https://roundcube.net

You can try it out at https://www.pikapods.com/apps#email to see if this works for you.


I’ve been a Migadu customer for a few years. It’s a simple company with simple pricing, and I couldn’t be happier.


Oh wow, that is a competitive rate vs fastmail!


> I probably would have gone Proton but they don't support forwarding.

They do on their paid plans. https://proton.me/support/email-forwarding


Ooh, looks like they added that in late 2023. Man. In August 2023 I actually migrated all of my email to Proton and was ready to go when I realized they didn't support forwarding. Thanks for letting me know.


In your Inbox,

or whole mailbox?


Inbox. They claim that the inbox is special in IMAP and it's hard to have a lot of messages there. 150K messages in the whole mailbox, I think. 25 years of email.


Unless I'm misunderstanding, treating your email in this way sounds like a crime against yourself and possibly humanity.


Apple's Airpods Max headphones appear to be the official uniform of University of California students. We've been visiting and I swear they outnumber normal headphones.


You will learn to think different like everyone else.


I went through several pairs of crappy wireless headphones until I gave in and got airpods, which just work.


Wired headphones also just work, don't need to be charged, and are much cheaper. AirPods are a strict downgrade from normal wired headphones, and it is insane to me that people are willing to pay for them.


Except when they get stuck in the treadmill and yank the cord out of the headphone. Wired headphones for working out and jogging are a nightmare.


For a year, and then the internal battery is useless.

They are engineered like consumables. Utter insanity. For over $100.


I've been wearing my AirPods pro since Covid with no issues.


Since less water would increase the detergent concentration, I was wondering if the opposite was the case. My family's old washer filled up the entire tub with water, so any detergent (and any pathogen, to be fair) would be quite diluted.

Short cycle length certainly makes sense to be correlated with pathogens. The lousy LG "TurboWash" only takes 28 minutes to do a full load of laundry but certainly doesn't get very much clean in that time.

I have to admit it was surprising that textiles have been identified as the source of hospital acquired infections. You'd think that even if the laundering didn't eliminate pathogens, it would greatly reduce them and make any clusters more diffuse.


As I understand, it's been identified as one possible vector, not conclusively proven to be the only (or even largest) source.


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

Search: