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

Interesting that models still evolve fast enough that dedicated model-specific hardware isn't a big contender right now. We're still seeing major scaling gains on mostly generic platforms.

google tpm, groq and cerebras needs yo be mentioned even if they are more general architecture optimized.

This is interesting. I would be curious to try a setup where I keep a local hybrid cache and transition blocks to deep storage for long-term archival via S3 rules.

Some napkin math suggests this could be a few dollars a month to keep a few TB of precious data nearline.

Restore costs are pricy but hopefully this is something that's only hit in case of true disaster. Are there any techniques for reducing egress on restore?


The easy way to achieve that, is using ZeroFS NBD server with ZFS L2ARC (L2ARC with local storage and “main” pool on ZeroFS).

Hopefully this is a wakeup call for anyone thinking that phone number validation is sufficient to prevent botting and fraud.

There's no such thing as (completely) "prevent", just substantially reduce by making it more expensive.

Which makes me wonder how much these guys spent per phone number. If you figure you can use a phone number maybe 2 or 3 times per service before it gets blocked they could create literally millions of accounts with this operation across various services. I don't know what the black market value for a Gmail, Facebook, Instagram, Reddit, HN, etc... account is, but I suspect it is more than the price of a SIM card.

The SIMs pictured in the article are from a carrier called MobileX so that’s the first clue to finding that out. Looks like a couple bucks is enough to get started, but a dealer on the take may have more attractive options and the ability to activate a lot of SIMs without the fuss of signing up accounts with email for each one.

About a $1 for a Gmail, IG etc these days.

I've used this before when I need to create an IG for a work project without wanting to link it to a personal number:

https://www.textverified.com/verifications


AliExpress has a beautiful 7" screen for peanuts. It has a surprisingly clear display and has so many uses:

https://youtu.be/LC3INaZVqFA?si=2BV5N3_7TtWPRlUj

It even has USB power and speakers.



I feel like the “bending ice makes electricity” bit is years, if not decades, old. Now I’m off to explore the rabbit holes and understand my own memory.

I searched as much as I could but couldn't find anything. Please let me know, I feel like my memory is faulty.

Poking at this dissertation (Continuum and Computational Modeling of Flexoelectricity by Sheng Mao):

https://repository.upenn.edu/server/api/core/bitstreams/3bc2...

> Flexoelectricity, ever since its discovery, has been regarded as an alternative of piezoelectricity at small scales. In fact, as early as the 1960s, Koehler et al. (1962), Turch´anyi, G. et al. (1973), Whitworth (1975) found that edge dislocations in centrosymmetric materials, such as sodium chloride, carry charge. Later, Perenko & Whitworth (1983) extended the observation to another kind of centrosymmetric material, ice. Piezoelectricity vanishes in these materials, therefore cannot be the source. Instead, a “pseudo-piezoelectric” effect was postulated by Evtushenko et al. (1987) for an explanation, which was later shown to be a result of flexoelectricity Mao & Purohit (2015).

Emphasis added.

So I think this was known but not fully understood by the time Perenko & Whitworth published Electric currents associated with dislocation motion in ice in 1983?

https://scholar.google.com/scholar?cluster=69134483967663101...

> In this paper we describe [an] experiment in which a small current is observed due to the movement of dislocations during plastic deformation [of ice].

The same authors of the paper TFA discusses published a preprint in 2022, which could also be what you're thinking of: https://arxiv.org/abs/2212.00323


Given the current err climate of thought purity, doesn't this seem a little too risky of a product to enable?

I feel like we are past that point now. The fact that AI will get things wrong has been normalized already


The kitty key protocol solves this, but your app and terminal need to support it. Many do.

This is the first time I've ever seen a macOS update and not seen a single feature worth bothering to upgrade over. Is there anything developer-facing? I don't use any Apple ecosystem stuff and this is all that AFAICT


There is support [1] for containers.

[1]: https://github.com/apple/container#requirements


Social media is 90% outrage farming by nation states and individuals with agendas. I think the experiment failed. Humans in large groups are so trivially manipulated by algorithms.

Our ape brains at large just can't deal with a firehose of manipulation. We're just giving bad actors a key to our subconscious to destroy the fabric of our civilization, and those bad actors are using it as much as they can.

This article is really missing the discussion on the fact that social media is far, far more inauthentic than real humans these days.


For me

Twitter: anger

Instagram: vanity

Facebook: pride

Of those, I hate anger the most so I am off twitter, even though it’s still a great place to overhear the latest in the business, geopolitics and tech zeitgeist. You literally hear about things before anyone else. But I decided that life is too short to be angry all the time.

I like instagram because it’s entertaining (and am of the age where I can ignore influencer antics). This is not true of younger folks for whom instagram is a mimetic model (there’s so much bad advice on it — especially on how to be a man or a woman today). As an older person, I want to tell young people: don’t get relationship advice off instagram. They’re so shallow and make you so entitled — “if your guy doesn’t do this, you’re too good for him” is pretty toxic.

Facebook used to be this political rage bait place but everyone’s left so I’m still hanging around with my geezer friends in their 40s. It’s become a place to share travel photos and pictures of kids.

Tiktok: I’m too old for Tiktok but I hear young people use it for news.

Google: I still use Google but young people are now increasingly abandoning Google and search on ChatGPT.


I like your summary, but I'd add...

Instagram & TikTok: bad (or even dangerous) medical and self-help advice


I agree, but we all know that political violence (or any radicalized violence) is not new. Social media, or any kind of mass communications, just makes it easier.


I think the outcome here is we tried nothing, and now conclude nothing works. Ok, we tried privacy breaking "protect the children" nonsense. We used to have a much more of public announcement/education system for aspects of society for civil education (School House Rock), for nutrition, for public health (vaccination). All of it has since been deemed frivolous public expenditure - maybe because it worked too well and the need became invisible.


Fully agreed.

Letting Big Social and governments sock it out to try to make this work doesn't seem like a winning proposition.

I really cherish Bluesky / atproto in huge part because it offers some chance for organic exploration, for individuals & orgs to test and try many ways to make social media work, to defang the Advanced Persistent Threats of the info-world. Without data that wouldn't necessarily count for much, but the firehose & backfill are also quite easy to get: humanity stands some chance, isn't quite in such a dark forest with the adversary here.


Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

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

Search: