sangnoir

Whatever cleaning they do is not effective, simply because it cannot scale with the sheer volumes if data they ingest. I had an LLM authoritatively give an incorrect answer, and when I followed up to the source, it was from a fanfic page.

Everyone ITT who's being told to give up because its hopeless to defend against AI scrapers - you're being propagandized, I won't speculate on why - but clearly this is an arms race with no clear winner yet. Defenders are free to use LLM to generate chaff.

The same LLMs tag are terrible at AI-generated-content detection? Randomly mangling words may be a trivially detectable strategy, so one should serve AI-scraper bots with LLM-generated doppelganger content instead. Even OpenAI gave up on its AI detection product

Amazon is famous for making losses after being publicly listed. Also, it was remiss of grandparent to not note that Amazon's losses were intentional for the sake of growth. OpenAI has no such excuse: their losses are just so that they stay in the game; if they attempted to turn profitable today, they'd be insolvent within months.

...and what a miserable bunch most of them were: 2 MB attachment limits and 10 MB mailbox quotas?! GMail had to drag all of them kicking and screaming into the 21st century when they debuted with a 1 GB mailbox quota.

The fediverse, like email and most other human systems that benefit from economies of scale will have a Pareto distribution with very large free providers > ISP / work and community instances > individual instances. Banking is decentralized and interoperable, this doesn't prevent banking giants from existing the world over - I expect them to exist, I also am in favor of smaller regional and local banks and credit unions.

1. https://www.litmus.com/email-client-market-share

2. https://www.oberlo.com/statistics/most-used-email-clients

Interestingly, the word "vagabond" is getting a rebound as "homeless" has become really unpopular lately, and as of 2022, "vagabond" saw more usage than "unhoused", which one might not have guessed given the bellyaching over other people's language.

1. https://books.google.com/ngrams/graph?content=vagabond%2Chom...

Sender-reputation management and anti-spam automation is far advanced and far more stringent for email than it is for Mastodon/the fediverse. Setting up a Mastodon instance is a matter of installing a package and going through a browser wizard comparable to setting up a Wordpress instance, not difficult at all.

The whole appeal of email-style decentralization is that even in those cases, Google can only block you from GMail - not email as a whole, which would be far more deleterious, and is the ever-present threat in non-decentralized platforms.

He was the opposite of narcissistic, he was honest to a fault, telling the truth even when it was unflattering to his public image: he freely admitted - and without direct prompting - in a Playboy interview that he had looked at a lot of women (other than his wife) with lust[1].

1. WARNING - site is on a Playboy domain - you may not want to open this on a work computer or have any anti-NSFW mononitoring of your browser activity. https://www.playboy.com/read/playboy-interview-jimmy-carter/

His Strava log clocks him at double that (34+ miles per day) over the last couple of days. So I'd say the average is not representative of his normal, on-the-road pace - which makes sense because he was detained for 3 weeks, visited extended family for part of the 517 days in addition to whatever breaks he took.

I cant be the only person who uses Quake-style terminals at fullscreen. The second part of your sentence is the crucial bit: the ability to instantly conjure a persistent terminal regardless of whatever else I have on screen.

I can't get the site to load

There was a "show hn" many months ago that did stylometry on HN commenters to show which accounts were most stylistically similar; I ran a throwaway account of mine through it, and it showed my account in the top 3 - which was impressive.

Having multiple accounts won't save you when your own word choice, grammar and style can uniquely identify you to anyone sufficiently motivated to link your disparate identities at any point in the future. The author even said their tool was rather basic; IIRC the basis was all pairs similarity on n-grams

As you noted, the first[1] batch of mass-media "influencers" were AM Radio hosts, then came cable infotainment , and now vloggers.

1. Perhaps preceded by syndicated newspaper columnists, but there was an editorial filter for those.

The same article also states that AMD provided custom bug-fixes written by Principle Engineers to address bugs in a benchmark - this is software that will only become part of the public release in 2 quarters. I ask again, do you think AMD will not expedite non-public bug-fixes for hyperscalers?

> You can't train with AMD, full stop, because their software stack is so buggy.

Point 7 from the article:

>> The MI300X has a lower total cost of ownership (TCO) compared to the H100/H200, but training performance per TCO is worse on the MI300X on public stable releases of AMD software. This changes if one uses custom development builds of AMD software.

I have an inkling that Meta does not obtain MI300 drivers from https://download.amd.com

[citation needed]

Do you honestly think the MI300 has show-stopper driver bugs, or that Meta/Amazon doesn't have a direct line to AMD engineers?

Would you say AMD is "shitting the bed" by not building it's own consoles too? You know AMD could build a kick-ass console since they are doing the heavy-lifting for the Playstation, and the XBox[1] , but AMD knows as much as anybody that they don't have the skills to wrangle studio relationships or figure out which games to finance. Instead, they lean hard in their HW skills and get Sony Entertainment/the Xbox division do what they do best.

1.and the Steam Deck, plus half a dozen Deck clones.

It's not a deflection, but a straightforward description of AMDs current top-down market strategy of partnering with big players instead of doubling down to have a great OOBE for consumers & others who don't order GPUs by the pallet. It's an honest reflection if their current core competencies, and the opportunity presented by Nvidia's margins.

They are going for a bang-for-buck right now aiming at data center workloads, and the hyperscalers care a lot about perf/$ than raw performance at. Hyperscalers are also more self-sufficient at software: they have entire teams working on PyTorch, Jax, and writing kernels.