swiftcoder

4 ms·

61.
by swiftcoder 15d ago
Having subsequently met a few folks who worked on Spore, I'm pretty sure they fully believed in the early vision they showed off a year or so before release
62.
by swiftcoder 15d ago
Smarter IDE folding, in effect? I'm always wondering what higher-level views of program flow we could have. Various tools have experimented with generating graphs and flow charts from code, or visual programming systems with nodes, but
63.
by swiftcoder 15d ago
Right, easily run 128-160 GB models, yes. Not easily run models much larger than that. Anything that doesn't fit in the 192 GB (including context, cache, etc) is going to have to be sparse/MoE, and the rather anaemic PCIE bandwidt
64.
by swiftcoder 15d ago
The return of Hong Kong to China is the first global news item I really remember from my childhood (the death of Princess Dianna shortly thereafter being the second). One really felt a shift in england that day: this then is the final nail
65.
by swiftcoder 15d ago
It's a very pretty effect.
66.
by swiftcoder 15d ago
> Not sure if you misunderstand what GPUs we're talking about, one RTX Pro 6000 has 96GB of VRAM. Right, but the top option here is a pair of RTX Pro 6000s, hence 192 GB of VRAM in total. Should be enough for a 128GB model plus co
67.
by swiftcoder 15d ago
> engagement falls by half without the algorithm Good? I honestly only have these apps to see what my friends posted. I know they aren't likely to do it, but it would be lovely to have the friends-only view back by default in every
68.
by swiftcoder 15d ago
> performance is more than just the aggregated memory speed of the hardware you have In other fields, sure, but for big LLMs it's a very significant part of the performance picture. That stack of Macs also is going to be able to nat
69.
by swiftcoder 15d ago
Are you sure about that? For the price of the top configuration here you can buy a 5-stack of 256GB Mac Studio M5 Ultras. This dual NVIDIA RTX PRO 6000 setup has 192GB of VRAM at 1.8 TB/s, versus the 5-stack of Macs that collectively h
70.
by swiftcoder 15d ago
I'd be interested to see how Reasonix stacks up here - they seem to have spent a lot of effort on tuning prefix cache reuse
71.
by swiftcoder 15d ago
$37,000 in GPUs. Man, I never expected another PC manufacturer to make Apple's top configuration look cheap by comparison.
72.
by swiftcoder 16d ago
> some use "not accredited" as a pejorative term. I am absolutely using "not accredited" in a pejorative sense here. That he publishes papers coauthored by a couple of philosophy professors at Oxford (all of whom have
73.
by swiftcoder 16d ago
Specifically, it joins a list of existing Tier 1 languages (C++, C#, and TypeScript), as "one of the best-supported languages for internal development at Microsoft"
74.
by swiftcoder 16d ago
> Obviously NMS's original marketing was bonkers and capitalised on saying technically true things in ways to mislead large numbers of people. I don't think they intentionally misled people. It's not that different to ho
75.
by swiftcoder 16d ago
> If you have access to a SOTA model without guardrails, provide a prompt that lets the agent come up with "creative" solutions to problems, and don't properly isolate it, they can end up inadvertently hacking 3rd party co
76.
by swiftcoder 16d ago
FB's main monorepo was named WWW, and you needed to refer to it every 5 minutes, so everyone pronounced it "dub-dub-dub". A surprisingly hard habit to break once you leave their employment
77.
by swiftcoder 16d ago
I guess the big question is whether they are going to open up rustc_codegen_utc for use outside of Microsoft?
78.
by swiftcoder 16d ago
They've recently added a ton of detail to the pricing page, but you'll have to calculate out tokens/month from that: https://opencode.ai/docs/go/#usage-limits
79.
by swiftcoder 16d ago
Yeah, I’ve rarely seen throttling unless fanning out to a ton of agents
80.
by swiftcoder 16d ago
The field was purely theoretical 20 years ago, and Yudkowsky is pretty much the dictionary definition of "not accredited"
81.
by swiftcoder 16d ago
OpenCode Go is currently running a 4x usage promo on DeepSeek v4.1 flash, not a bad way to get your feet wet (even if their cache hit prices are probably still very sub-optimal)
82.
by swiftcoder 16d ago
I've literally only done this once per computer I use the web interface on, and it's always stuck thereafter. I can't speak to the mobile app
83.
by swiftcoder 16d ago
> scientists who have spent their lives studying this Please point me to one actual accredited scientist who has spent a lifetime studying AI alignment? Pretty much this whole field is only 5 years old
84.
by swiftcoder 16d ago
> Why focus on the style of writing instead of the subject matter? Because just as with human writing, a poor writing style obscures the topic where it should illuminate
85.
by swiftcoder 16d ago
For some reason I get a black screen in Safari. Renders ok in Brave on the same MacBook Pro.
86.
by swiftcoder 16d ago
Here I thought we were discussing a return of the Aliens vs Predator franchise
87.
by swiftcoder 17d ago
For sure. Even as someone who bounced off the game several times, the scope of their post-launch additions are truly impressive, even matching up against a lot of the subscription games (whose whole deal is regular content drops)
88.
by swiftcoder 17d ago
> I feel exactly the same. There has to be a way to do procedural generation in a way that doesn't feel "sparse" A good start is not to try and build an entire galaxy. Procgen that focuses on the fine detail of a relativel
89.
by swiftcoder 17d ago
huh. mine only does this on the first turn on a new computer. Once I've told it once it seems to be entirely sticky on that device from then on
90.
by swiftcoder 17d ago
Nah, we're long past the point where that would make a difference - if they could have done so effectively, they would have before K3 and GLM 5.3 were nipping at their heels...

More ›