nyx

It's 200GbE. Assuming you could actually fully saturate both of those links, it would only take a couple of hours to copy off the entire capacity. I think you'd more likely be capped by the write speed of the destination storage?

Huh, didn't know about "ChatGPT-User"... I've always been a fan of a slightly blacker-hat approach to this problem: rather than block those user agents outright, wouldn't it be way more fun to serve them nonsense or irrelevant garbage-filled versions of your pages?

Counterpoint: this YouTube rant by an animation person called Noodle is a pretty good overview of why frame interpolation sucks. https://www.youtube.com/watch?v=_KRb_qV9P4g

Basically, low FPS can be a stylistic choice, and making up new frames at playback time often completely butchers some of the nuances present in good animation.

VPP is really neat tech. I recently worked on a product that employed it, and it was impressive to see a commodity low-power CPU pushing tens of gigabits of traffic.

Oh man, on first reading I definitely thought it was saying "with experimental technologies, Linux can drop all traffic", which would have been a hilarious dig... but I think "drop all incoming packets" is a useful benchmark for evaluating overhead, like doing conntrack for every packet that hits then immediately discarding it.

Ahh... I miss n-gate's webshit weekly. You know he'd have been all over this thread.

Omegle (business model: "Uber for child grooming")

I genuinely thought this was going to be satire, based on just the headline... then I clicked and read the first paragraph, and STILL thought it was satire.

Then I saw the link to the earnings release and figured out it's real. Where else but tech is "things are great and we're firing 10% of our workforce" a normal announcement?

> rebellious streak

Translation: says the N word?

Sure, I could see the value in an LLM that knows a lot about current events or whatever, but I'm wondering if this is going to be an LLM that behaves like the median post-acquisition Twitter user.

This is a problem they're working on solving, I think. The hospital group I've been using in the Portland area recently announced that patient-initiated messages that take more than a few minutes, and require e.g. digging through a chart, are billable to insurance: https://www.legacyhealth.org/messages

MyHealth is just what this hospital system calls their patient-facing Epic portal.

Maybe a little off-topic, but I always wonder if the folks who work on stuff like this ever consider that they might be the bad guys.

Like, sure, maybe the people you're trying to send email to are already customers, or they're otherwise really interested in hearing what you have to say, and maybe that's what they all tell themselves. But "tips for disguising your unsolicited commercial email as legitimate communication so you can succeed in shoving it into people's inboxes" feels like it'd take a lot of reality distortion to not feel guilty about.

Am I just naive? A Hacker News sheep living on grass while the wolves feast on red meat? Is there such a thing as being a conscientious objector for spamfucking people with trashy ads and sales pitches?

I've seen a bit of discussion lately about EPSS[0], which is a model that tries to take more practical metrics about any given CVE (including pieces of the CVSS, but also things like whether there's a GitHub PoC or if it's up on ExploitDB) and return a real-world exploitability score out of 100.

I can't help but imagine something like this is where we're headed, given how counterproductive it can be to make decisions based on CVSS scores...

[0]: https://www.first.org/epss/model

Wait... so this isn't satire? I read all the way through expecting a reveal, but when I got to the bottom all I found was a link to other chapters of "How to Productize Yourself"... which, again, I thought was a joke but upon clicking that I got a list of articles like this one and zero indication that the author is joking.

I know this person is capable of sarcasm and irony and so on, because I remember seeing their "how to be a -10x engineer" article[0] before... if they're intentionally riding the line of believability here, then props to them, because this is hilarious and definitely sounds like something in the top quartile of unhinged LinkedIn-influencer-posting.

> Create arbitrary rules and rituals. Ask your clients to participate in your personal process. Lead people on journeys.

What's wrong with unassuming and consistent competence? Has the author ever actually worked with another one of these high-output culture-hacker people? It's possible to leave a great impression with nothing but technical excellence and solid communication, and I'd personally take that any day over someone who tries to turn a client meeting into some kind of spooky grindset seance.

[0] https://taylor.town/-10x

Are you thinking of SomaFM's SF 10-33? https://somafm.com/sf1033/

The purpose of this story is for the author to recount their experience in an entertaining way that's pleasant to read. Why would you want to use an LLM to summarize it? That'd be like finding an interesting-looking novel to read at the bookstore, then buying the Cliff's Notes instead.

I've thus far kept myself mostly isolated from and ambivalent about the GPT hype, but this comment, even if it's somewhat tongue-in-cheek, really says a lot about where the technology is going to take us as a species. I'm not thrilled about it.

Yeah, that's completely right, but as you said in another comment, there's some real bullshit in the FMVSS that to me, a complete layperson, seems like it has the sole effect of blocking the certification of perfectly safe and clean modern vehicles from other advanced nations.

Of course we shouldn't be allowing people to import some pollution-spewing deathtrap that doesn't have seatbelts--the FMVSS regulations do exist for a reason--but I think we should be taking a more critical look at our regulations, especially as compared to other places at the same socioeconomic level.

The 25-year import rule here, which bans Americans from importing vehicles from other countries unless they're 25+ years old, is just awful for those of us who would be interested in driving small, efficient foreign cars.

It would be great to live completely car-free, but absent major changes to how we plan our cities, it's just a sad reality that cars are a necessary ingredient to life in the vast majority of America. To cope with this, I'd love to be able to import a kei car or van from Japan, or micro-sized European city cars, or even some of the very small EV city cars that we see in China... but I just can't, unless I want an overpriced pile of scrap from the 1990s.

It's all so much worse when you realize that the 25-year rule is a holdover from a grey-market import scare of the mid-1980s[0]: European carmakers, namely Mercedes, BMW, and Porsche, were having trouble in the US with people importing European models of their cars. There were some valid concerns around inconsistent modifications for US safety standards, but the main issue was clearly that these grey-market imports were cheaper than buying a US model from a dealer, so profits were being missed. Instead of fixing the pricing discrepancy, they just successfully lobbied the government to enact this draconian 25-year ban, and so to this day I can't have a 2020s Japanese kei car shipped to a US port at my expense because it'd be illegal to register it.

[0] https://jalopnik.com/the-25-year-import-rules-history-is-mor...

I don't know and don't want to speculate, but I do hope it's just a fun indie project with the goal of artistic expression, and that there's no parasitic "how-can-we-monetize-this" ulterior motive.

> it was because local people decided it was in their interests to pay a dollar less at Walmart

Well... isn't it? I'm not sure I like the implication behind how you've worded this. It's easy, as somebody fat with tech money, in the political know and with the benefit of hindsight, to call out those small-town rubes for their myopia, to say "but of course Walmart is a monopolist, of course the local grocer will suffer if you give Walmart your business; markets are zero-sum, isn't it obvious?"

But I don't think that's really fair. If I get paid a few bucks an hour in middle America and every dollar counts, obviously I'm going to go for the Great Value breakfast cereal or whatever: I don't really have the luxuries that would permit me to vote with my wallet.

Ahh, /etc/machine-id, we meet again... Last time I met you, you were causing my cloned VM to get the same DHCP address as its original and blow up my VM network, despite libvirt's virt-clone utility properly randomizing vNIC MAC addresses, because netplan defaults to using you as the DHCP client identifier. If only we could meet under more pleasant circumstances.

Yeah, the text to the right of the input is explicit about this:

> Currently, there are 11817 unicode character glyphs in the database. Japanese, Korean and Chinese characters are currently not supported.

Directly connecting these blinds to my Home Assistant instance via the ZHA integration and a HUSBZB-1 Zigbee/ZWave stick[0] works pretty well. I've got 8 of them around my house, and some automations to open/close with the sun, etc.

Only small issues I've identified are that 1) Zigbee meshing doesn't seem to behave with these, so you'll have to pair the blinds directly to your hub. This is fine, as long as your place is small enough for there to be no signal quality issues. And 2) if Home Assistant instructs the blinds to open to a specific position (i.e. "80% closed") rather than fully open or fully closed, it can take a really long time for the blinds to update the hub with their state. This means they show as "opening" in the HA UI even though they're finished moving.

[0] https://www.amazon.com/dp/B01GJ826F8

When I got my new HVAC system installed a few months ago (heat pump, woohoo!), they needed to upgrade my thermostat, so they gave me a Honeywell VisionPro 8000, which is a "smart" thermostat but, like your T9 (I think), demands WiFi and some garbage "cloud" service to access the connected features.

I replaced it with a T6 Pro, a thermostat which is by all other metrics a downgrade, but which has a single killer feature: Z-Wave. No phoning home, no janky embedded devices on my WiFi network, just a solid integration with Home Assistant that falls back to a regular thermostat if my HA server is down.

Seems like Honeywell is moving away from making Z-Wave versions of their products, because I think most consumers are generally content to tunnel their request through some shitty cloud thing every time they want to turn the heat up... but for those of us who are a little more ideologically rigid about keeping their stuff LAN-only, would definitely recommend grabbing a T6 Pro while they're out there.

I always wonder what kinds of people benefit from writeups like this. Maybe I'm deceiving myself, but I feel as though I'm a pretty good conversationalist, despite not having much of an in-person social life in my formative years. The suggestions in articles like this one always seem pretty obvious to me.

Are there people out there who read this and realize "oh crap, I interrupt too much," or "huh, I should try to be more aware of the other person's interest level before infodumping on them"? My hunch is that people who might genuinely benefit from these suggestions are, to put it bluntly, generally too poorly socialized to even recognize that their behavior could use the adjustment.

A demographic I could see benefiting more from this kind of thing are the obsessive-optimizer class of folks who interact with others in a cynical, How-to-Win-Friends kind of way, and who see social interactions with fellow humans as something to be min-maxed. I'm sure those people are over-represented on orangesite, so this might be controversial, but I don't think that doing an intentional post-mortem on every social interaction you have is any way to live.

Handled my puerile test input pretty well, including photo: https://i.imgur.com/x60ZuDx.png

This doesn't seem useful beyond serving as a tech demo and shill post for OP's employer's subscription service. A lot of the nuance in recipes is lost when you programmatically clobber them into a format uniform enough to train a model with, and it seems like in this case it can either regurgitate existing recipes, or invent dodgy unappetizing ones.

How does this even happen? Is there some system automatically blasting out promotions based on a list of notable calendar dates?

Can we expect to get KFC coupons on September 11? How about a free side of mashed potatoes commemorating the Watts riots, or 2 for 1 meal combos on the anniversary of the Chernobyl disaster? Reach out, KFC marketing people--I'd be happy to come on board as a consultant.

And the email purported to have been written by Andy Stone (who does not seem to be Indian) is rife with indicators that it was written by an Indian English speaker:

  * How the hell <report> got leaked? (non-standard conjugation)
  * Why didn't anyone of you bother... ("anyone" for "any one")
  * Put <names> on watchlist. (dropped article)
  * Send me ... for last one month. (dropped article)

I think of the t motion as "to", and the f motion as "frough" (rhymes with through: "delete frough the second t").

The scan that indicated the 80% reduction was after just a couple of weeks, if I recall correctly.

Since you're talking about immunotherapy not typically being a first-line treatment, I'll share a morbidly interesting fact that underscores some of the, er... quirks of the US medical system. The oncologist treating my relative initially staged my relative's cancer in the electronic health records as stage IV, despite no evidence of metastasis (the usual criterion)--he explained that he did this specifically in order to pursue first-line Keytruda (which is indicated for stage IV but not stage IIIC) and have it be covered by insurance.

Thanks for sharing your story and treatment details. I should add that my relative's cancer has no particularly interesting mutations, but does have a high PD-L1 expression, which from what I can tell is the reason Keytruda was more likely to work for their situation.

> that was geared towards my specific mutation

This is actually one of the things that gives me hope for the future: genetic testing on a tissue sample of the patient's cancer is standard, and for many of the specific oncogenes that we know about, there exist therapies targeted to those specific mutations: https://www.cancer.org/cancer/lung-cancer/treating-non-small...

One thing I've learned is that even in the face of good news, cancer is a horrible time, and I wouldn't wish it on anyone. But I'm likewise glad to hear about your results from the latest treatment, and hope things stay as positive as they can for you.

I agree that my original comment is very optimistic--for what it's worth, Keytruda has thus far been very effective in the case I'm talking about, so I have some bias here. I concede that even in my "double the effectiveness" example, we're talking about doubling something like a 20% 5-year OS.

By "around the corner" I'm really talking about, like, 20-30 years out, which I think is fairly soon in terms of cancer treatment progress. You're right that it reads like a leap in the context of my comment.