jsheard

33 ms·

811.
by jsheard 1y ago
I wish I could be so optimistic that there is no private information published unintentionally or maliciously on the open web where crawlers can find it. (and as diggan said, the web isn't the only source they use anyway. who knows wha
812.
by jsheard 1y ago
Considering every AI company is hemorrhaging money with no end in sight, that doesn't bode well, does it?
813.
by jsheard 1y ago
Why single out user data specifically? Most of the data Anthropic and co train on was just scooped up from wherever with zero consent, not even the courtesy of a buried TOS clause, and their users were always implicitly fine with that. Fo
814.
by jsheard 1y ago
In theory a CPU with SMT could still trigger this bug, because not every core necessarily has to have SMT. Intel made some chips that combined performance cores with SMT and efficiency cores without SMT, so if they had an odd number of E-
815.
by jsheard 1y ago
The fun thing about those is they were physically quad cores with one core disabled, which may or may not have been defective, so if you were lucky you could unlock it and get a bonus core for free.
816.
by jsheard 1y ago
I think we'll have to agree to disagree about well done 3D renders not feeling real. Movie studios still regularly underplay how much CGI they use for marketing purposes, and get away with it, because the CGI looks so utterly real that
817.
by jsheard 1y ago
You're comparing apples to oranges, holding up today's practical real-time rendering techniques against a hypothetical future neural method that runs many orders of magnitude faster than anything available today, and solves th
818.
by jsheard 1y ago
AMDs desktop chips don't have distinct P and E cores, they're all P cores. AMD do have an E core design but it's currently only used in mobile and server parts.
819.
by jsheard 1y ago
Modern silicon is so dense and heats up so fast that throttling is easier said than done. I think they have to model and predict the thermals ahead of time nowadays, because by the time they could react to a temp sensor alone, the chip migh
820.
by jsheard 1y ago
This is what Zen5 looks like under the IHS: https://i.imgur.com/j85YUzX.jpeg Everything is offset towards one side and the two CPU core clusters are way towards the edge, offset cooling makes sense regardless of usage.
821.
by jsheard 1y ago
You know Llama is a Meta trademark, right? If you want to compete with the big labs then step one should be to avoid handing them a loaded gun to shoot you down with.
822.
by jsheard 1y ago
Valves big misstep with the Steam Machines was that they expected developers to port their games over to Linux natively, on their own dime. Needless to say that didn't end up happening at any significant scale, so when they resurrected
823.
by jsheard 1y ago
What happened recently to bring this on? MediaLab acquired Imgur years ago.
824.
by jsheard 1y ago
It's not looking good so far. When OpenAI introduced product searches back in April I tried running one of their own example queries from the announcement post, and it obliviously cited "reviews" and "recommendations&quo
825.
by jsheard 1y ago
Not the cheaters themselves AFAIK, but game companies have successfully gone after cheat companies. https://www.ign.com/articles/call-of-duty-cheat-maker-ordere...
826.
by jsheard 1y ago
Microsoft is paying top dollar for MarkMonitor, aren't they supposed to proactively register obvious typos so this kind of thing doesn't happen to their clients?
827.
by jsheard 1y ago
It's even recommended when building out a CCTV system with cheap Chinese IP cameras that like to phone home all the time. Stick 'em on a VLAN which can't access anything besides your local NVR.
828.
by jsheard 1y ago
Yep, and that applies to Valve at two levels because Steam dwarfs the ROI of their games, and their forever-games like Counter Strike dwarf the ROI of any singleplayer game they'd ever be able to make. It's a miracle they even got
829.
by jsheard 1y ago
Dota 2 is of note, but that was a pretty safe project being a straight remake of an already popular mod with the mods developer at the helm. Their major titles since 2012 were: CS:GO (remake of a remake) Dota 2 (remake) Artifact (fl
830.
by jsheard 1y ago
https://www.nvidia.com/en-us/data-center/rtx-pro-6000-blackw... The workstation versions are fine if you're running one or maybe two cards with an airflow gap between them, but if you pack four of them right
831.
by jsheard 1y ago
https://www.youtube.com/watch?v=mvBeCSaaDxA What did you say?? I can't hear you over these blowers!!
832.
by jsheard 1y ago
The technical expertise of... buying four of the fanciest Nvidia GPU and plugging them into an off-the-shelf motherboard in an off-the-shelf chassis? A serious attempt at this kind of build would use the server variant of the RTX6000 and cu
833.
by jsheard 1y ago
The Xbox Series S only uses about 80w under full load, and that's built on TSMCs old 7nm process. Apple's bleeding edge 3nm silicon should be able to do even more with a similar power envelope.
834.
by jsheard 1y ago
Sure, but we were talking about high-speed digital cables, not audio cables. When you're pushing 48gbps over copper (as in HDMI 2.1) the cable construction and connection integrity absolutely does matter, older HDMI cables don't w
835.
by jsheard 1y ago
Gold plating electrical contacts does at least do something useful though, it helps to prevent oxidization/corrosion. A better analogy would be gold plated TOSLINK cables, which unfortunately do exist.
836.
by jsheard 1y ago
I suspect Reddit would try to lean on the catch-all ass-covering clauses that every social network already had long before AI data licensing deals were on the table. Such as this one from Reddits TOS circa 2018: > By submitting user con
837.
by jsheard 1y ago
I'm sorry, but if you give up on something you would "love to use" just because LLMs are unable to oneshot it then you might be a bit too dependent on AI.
838.
by jsheard 1y ago
Are you getting the "access denied" page, or an infinite challenge loop?
839.
by jsheard 1y ago
I think there's probably a distinction to be made between deliberate, careful use of synthetic data, as opposed to blindly scraping 1PB of LLM generated SEO spam and force-feeding it into a new model. Maybe the former is useful, but th
840.
by jsheard 1y ago
The world is not prepared for the mental gymnastics that OpenAI/Google/etc will employ to defend their copyright if their big models ever get leaked.

More ›