kardos

7 ms·

31.
by kardos 2y ago
> A civilian could easily buy a small fleet of drones, equip them with small IEDs or sarin gas, and fly them into otherwise secured areas with large crowds. Easy? Is sarin gas freely available at big-box retailers?
32.
by kardos 2y ago
1) virus database gets updated, what was written virus-free with the previous database may not be virus-free on the current database. 2) removable storage devices 3) the system drive is not controlled during reboots You could imagine buildi
33.
by kardos 2y ago
> Well, keep in mind that the full linking step has to be done at the end of an incremental build. Does it have to be that way? For an incremental build, particularly in the "not much has changed except some function contents"
34.
by kardos 2y ago
It must mean cooling significantly below the target temperature, and thus wasting power to do it
35.
by kardos 2y ago
Do the AI PCs do the work locally or transmit everything remotely for processing like chatgpt?
36.
by kardos 2y ago
Could this be done with an IDE plugin (for Pycharm or whatnot)? It is tedious to go though code and carefully annotate each variable and run so many compile test, read error, revise, cycles. It should be possible on a per function basis to
37.
by kardos 2y ago
If the goal is to produce a LLM-like interface that generates correct output, then sure, it's not cheating..... but is it really a data-driven LLM at that point? If the LLM amounts to a chat-frontend that calls a host of human-prepare
38.
by kardos 2y ago
Thank you!
39.
by kardos 2y ago
Does it work with Intel fortran-compiled code?
40.
by kardos 2y ago
Sorting is costly though! This summing problem is solved by xsum: https://gitlab.com/radfordneal/xsum There is room for more work here to make it faster or use less memory though.
41.
by kardos 2y ago
Possible. I'm not posing that tar replaces sqlite here, just that tar should achieve the same space savings and syscall savings
42.
by kardos 2y ago
If the speedup is due to fewer open/close syscalls and space savings due to avoiding block padding, tar would achieve the same yes?
43.
by kardos 2y ago
Indeed it does not seem idempotent. If you move A to B and insert a new A, then rerunning the patch would yield the new A at B. I suppose a subset of this is idempotent though. Broadly this seems fraught with peril. It sounds like edge case
44.
by kardos 2y ago
> Yes, to be commercialized and privatized by the author. privatize: "transfer (a business, industry, or service) from public to private ownership and control." The outcome here does not include privatizing the scans by the aut
45.
by kardos 2y ago
There is no privatization here (moving the scans from public domain to private), the author is seeking the opposite, shifting the scans to the public unencumbered.
46.
by kardos 2y ago
> mine is a very basic cheap model with no cameras or wireless connectivity. What brand is it? So many these days have both cameras and wireless
47.
by kardos 2y ago
> "Are there any dishes you're out of, if we swing by in an hour or two?" > Full knowledge is easy enough for 90% success. Just record every human to human conversation for a month and use a RAG or fine tune the model o
48.
by kardos 2y ago
On the fingerprinting noise: this sounds like a job for FEC [1]. It would increase the size but allow using the Canvas API. I don't know if this would solve the flicker though (not a front end expert here) Also, it's a long shot,
49.
by kardos 2y ago
My takeaway from the linked post is that the author is more concerned with floating point invariance across platforms than speed (although improvements in speed are of course welcome). If the data confined to a certain range of exponents, o
50.
by kardos 2y ago
Exact floating point accumulating is more or less solved with xsum [1] -- would it work in this context? [1] https://gitlab.com/radfordneal/xsum
51.
by kardos 2y ago
> We need the option to view comments chronologically You might like this then: https://hckrnews.com/
52.
by kardos 2y ago
> Number of Pages > CCXXXI, 4622 Taf.
53.
by kardos 2y ago
You'd need geofencing of mobiles to be precise and accurate to about a meter for that to work. Is that the case?
54.
by kardos 2y ago
That is possible. Another possibility is that they are gradually watering it down in order to segment the market in the future
55.
by kardos 2y ago
FWIW, there seems to be a less-abandoned fork here: https://github.com/jazzband/docopt-ng I'm sticking with argparse though
56.
by kardos 2y ago
Yeah..... it seems like it would be fragile and require lots of iterating to converge on a help doc that is both "pretty" and correctly parsed by docopt
57.
by kardos 2y ago
That seems to be an abandoned project
58.
by kardos 2y ago
Maybe the CPython JIT can make use of it. That might move the needle past 1%...
59.
by kardos 2y ago
The limitations of the non-physics-based models is under rapid exploration and the bounds of their utility will be a lot more clear in a year or two. For now, they seem to outperform physics-based models in larger scales. The future may be
60.
by kardos 2y ago
> It was eating his lunch, so to speak How does that work. A tapeworm is pretty small relative to your brother (eg by mass), one would think it's caloric needs are similarly small and it would just use a small fraction of his food.

More ›