Presently, just Anthropic and DeepSeek.
177 karma · joined May 7, 2020
I'd say the biggest problem with the NPU is that you can only use it from Core ML. Even MLX can't access it it!
As you say the big world-changing LLMs are scaling up, not down. At the same time (at least so far) LLM usage is intermittent - we want to consume thousands of tokens in seconds, but a couple of times a minute. That's a client-server timesharing model for as long as the compute and memory demand can't fit on a laptop.
As I understand it:
- Anything which is breaking-changed in any way for 0.3.0 (e.g. in the article a type change from i32 to u32) is the exact same code in 0.2.1 as it was in 0.2.0.
- Anything which stays the same for 0.3.0 is re-exported in 0.2.1 (from 0.3.0). The code is replaced with a `pub use ...;`
- Anything which moves around in the submodule hierarchy (but is otherwise unchanged) in 0.3.0 is re-exported in 0.2.1, from where it used to be.
If crate Baz depends on something in Foo 0.2.x which changes in Foo 0.3.0, then when (if!) Baz updates to Foo 0.3.0 it will need to deal with that. But Baz can upgrade to Foo 0.2.1 without concern, and any types unchanged in Foo 0.3.0 will be interoperable.
Always funny to see 240V power described as exotic (or at least exotic-adjacent). It's the standard across practically the entire world except North and Central America!
But the US has a very powerful executive (the President) which sidesteps the problem. The US House of Reps could be multi party PR without any big issues (well, they might need to switch to a "last year's budget is this year's budge unless we vote to change it" model).
FYI, as of late 2021, you can opt into strictness.
The current plan is a temporary pool in a new inner-city arena (15k seats) for the major swimming events, with diving, artistic swim and water polo prelims at an existing (to be refurbished) sports centre.
Inside this zip file there's a number of things:
- three preview JPGs at varying sizes (roughly 50px, 250px, 1000px) - a Metadata folder with a couple of plists and a UUID file - an Index folder which contains a bunch of .iwa files which appear to be some Apple iWork suite specific thing. I'm not sure if there's a real file format there or if they all have a different format keyed on the file name. If you run `strings` on one of them you could plausibly recover document text.
For majors the benefit is clear: you don't get spoiled.
For minors, you don't have to overcome the barrier of being a spoiler if the race looks remotely close.
And of course voters are not disadvantaged just for having two acceptable choices rather than one.
America is the land of the lawsuit right? Surely someone - or the state legislature - could take action on safety grounds alone to ensure a reasonable minimum duration (adjusted for the speed limit) for a yellow light.
And it's not like speed traps, where the posted limits are well below the design speed of the road. A red light is a red light!
fn: We can argue the toss around discriminatory placement of cameras e.g. there's probably more of them in neighbourhoods with certain demographics, but there's a simple solution here: every signalised intersection getting a camera, or at least the top half by traffic volume.
e.g. in Mission Control right now I have "Desktop 1", "Desktop 2", and full-screened iTerm2 on my external monitor, plus "Desktop 3" and about 6 other fullscreen apps on my laptop.
I have somewhat semantic meaning for what each of those three numbered desktops are, but it'd be nice to _name_ them.
The technical term near me seems to be something like "delineator posts" (or just "orange posts" after the colouring) and I think that's pretty reasonable. As you say, they don't provide any protection against a car or truck, but they do signal where not to be a bit better.
More seriously, I'm sure hydrogen will make a valuable contribution to the stationary grid-level energy storage situation. I'm not convinced about other applications.
There's DB internals/performance course which I'm a TA for and we lean heavily on being able to force Oracle's rule-based or cost-based optimisers --- because then we can get the students to analyse and compare the plans.
Making money in the Oracle ecosystem without Oracle getting a cut, presumably :)
In Australia we usually think about undergrad competitiveness in terms of the minimum ATAR rank of last years' admissions, right?
But you could also think about competitiveness in terms of admission fraction.
And I think those two metrics can be consistent if a very popular degree has a disproportionately low fraction of high-ATAR applications.
The admins of the groups are responsible for managing the discussion in groups including frontline anti-spam against the porn bots and the shoe bots and whatever. So it's semi-distributed moderation. (Assuming group admins care, but if they don't then users just abandon the group.)
Compared to e.g. Reddit, FB groups aren't as world readable or writable as subreddits are so you don't really get that classic inter-subreddit drama.
Where global moderation is still needed from FB's perspective is when group activities start spilling outside of that group, e.g. because they're coordinating extremist activity.
This probably won't be the result of user reports though. For privacy and scalability that implies an automated system which (hopefully reliably but only rarely) flags posts to a human working for FB.