blueblimp

11 ms·

61.
Puzz.link: a database of Nikoli/WPF-style logic puzzles(puzz.link)
1 points by blueblimp 4y ago|0 comments
62.
by blueblimp 4y ago
That's a good point: the effectiveness at detecting AI generation is probably going to depend strongly on the length of the text.
63.
by blueblimp 4y ago
> true positive rate of only 26% and a false positive rate of 9% That's uninformative enough that I'm surprised they launched this publicly at all.
64.
by blueblimp 4y ago
This smells like a fake story created for trolling purposes.
65.
by blueblimp 4y ago
For PM2.5, the Qingping Air Monitor (or the Lite version if you prefer that) is good. The review I've read found it has decent accuracy, and, though I haven't tested it rigorously myself, the readings seem reasonable in my experie
66.
by blueblimp 4y ago
> Hell, just announce "you must have a Twitter Blue subscription to use a third party client" and that might have actually gone well for them. I've been wondering too why they didn't do this. That would at least get t
67.
by blueblimp 4y ago
> I'm sure some folks are thinking "what if you just limit yourself to 30 minutes a day?" An alternative to this that I've found more effective is to prescribe which modes of interaction are allowed. For example, in t
68.
by blueblimp 4y ago
Interesting point. That's definitely visible in the overall art style. But at least regarding the edges and corners of an isometric rectangular cuboid tile, I'm not convinced that there's a meaningful difference between pixel
69.
by blueblimp 4y ago
This seems to be just one approach to the problem. To my taste, the way they're drawing the vertical lines is ugly (because of the asymmetry), so it made me curious how notable professional isometric pixel art has solved the problem. I
70.
by blueblimp 4y ago
I spent a minute thinking ROTOR was impossible until I realized the asymmetry in the ruleset. Neat puzzle.
71.
by blueblimp 4y ago
That example reminded me of dreaming. Whatever random elements come up in a dream, my mind finds a way to rationalize them as making sense. (At least until I awake and realize that it was nonsense.)
72.
by blueblimp 4y ago
Past generations got rid of the wolves for a reason. Although it's nice for wildlife to exist somewhere, it doesn't belong in close contact with humans. The desire to bring wildlife to where people live reminds me of anti-vax in t
73.
by blueblimp 4y ago
I'd read a lot of coverage of this fiasco already, but this post was helpful for clarifying the duality aspect and the philosophical issues around what's real vs what's a simulation. I disagree a bit on tone though: though it
74.
by blueblimp 4y ago
> "It was, I thought, a dream fulfilled," Chang said. "But it [the first plant] ran into cost problems. We ran into people problems, we ran into cultural problems. The dream fulfilled became a nightmare fulfilled. It took
75.
by blueblimp 4y ago
Superficially, they look good. That's what's new about 2022 AI image generation, and it's legitimately impressive. But as I've seen more of these images (and especially used the tools myself) and taken a closer look, I&#
76.
by blueblimp 4y ago
> With the advent of generative AI art, it seems immoral from a fiduciary responsibility point of a view that an art director doesn't train an AI model on their illustrator's art before laying them off. If they do that, the qua
77.
by blueblimp 4y ago
Imagine if Photoshop were deemed somehow unsafe without a mechanism to prevent the user from creating NSFW images. The panic around image generation models is absurd.
78.
by blueblimp 4y ago
Plausibly OpenAI/etc. trying to get a competitor shut down.
79.
by blueblimp 4y ago
Why is this random person's speculation linked as if it has credibility?
80.
by blueblimp 4y ago
In the Salesforce example, it's modifying database contents. Suppose the model misunderstands your request and modifies data in an unintended way (e.g. adding garbage data or, worse, deleting data). What's the recovery plan?
81.
by blueblimp 4y ago
"The herbivore Ectoconus (a member of the Periptychidae, which may be related to living hoofed mammals, the ungulates) reached about 100kg (220lb) within a few hundred thousand years of the extinction. In geological time, that's t
82.
by blueblimp 4y ago
It's pretty wild that the author gets a somewhat-usable result merely by few-shot prompting OpenAI Codex.
83.
by blueblimp 4y ago
> And you can't just beam the waste heat from every process away from the surface of the earth, that violates the second law of thermodynamics. Surely this can't be right, because if it were, air conditioning (the normal everyd
84.
by blueblimp 4y ago
Most implementations probably use algorithmically generated puzzles, which are less fun than human-created puzzles.
85.
by blueblimp 4y ago
Thanks, it's been a while since the last time I played GMPuzzles... clearly I should do it again.
86.
by blueblimp 4y ago
I'm also a fan of Nikoli-style puzzles. (My favorite is Nurikabe.) Nikoli themselves used to offer an iPhone app with lots of good ones, but it's sadly unavailable now. https://puzz.link/db/ is a great source
87.
by blueblimp 4y ago
I liked Infinifactory the most because being in 3D space maximizes the difference from normal programming.
88.
by blueblimp 4y ago
I highly recommend Redactle as a similar concept that is more fun: https://www.redactle.com/
89.
by blueblimp 4y ago
I use Microsoft Authenticator on iOS, and it doesn't use location. (I didn't even need to deny it--it didn't ask for it.)
90.
by blueblimp 5y ago
Thanks for finding a link. As a civilian example, this document proposes applying it for medical staff dealing with COVID: https://www.wrair.army.mil/sites/default/files/2020-04/iCOVE... .

More ›