nextos

28 ms·

451.
by nextos 2y ago
NeXT was great and Mac OS X was also nice and had a lovely indie and boutique app ecosystem during the mid-to-late 2000s. Sadly, iOS stole the focus. However, the OP argues Linux usability is bad, which I think is an outdated POV. It really
452.
by nextos 2y ago
That is a good question. They sell those CPUs to industrial clients. Note cases can be configured to be completely sealed for high-dust environments, and it is also possible to get industrial motherboards with connections that no regular us
453.
by nextos 2y ago
Cirrus7 is expensive because you are paying for a very high quality machined chassis & case that act as a massive fanless heatsink. Those alone are pretty costly. The price cannot be compared with cheap NUC clones and mini PCs, nor with
454.
by nextos 2y ago
Minix if your budget is $200-400. Cirrus7 if your budget is $500-1000. There are many other manufacturers. I am biased towards fanless builds, like those two. NUCs are also a great option, especially if you replace the case with a fanless A
455.
by nextos 2y ago
Exactly, the M2 Ultra is competitive for local inference use cases given the 800 GB/s bandwith and a relatively low cost and energy efficiency. The M4 Pro in the Mini has a bandwidth of 273 GB/s, which is probably less appealing.
456.
by nextos 2y ago
For inference, Apple chips are great due to a high memory bandwidth. Mac Studio is a popular choice in the local Llama community for this particular reason. It's a cost effective option if you need a lot of memory plus a high bandwidth
457.
by nextos 2y ago
New Mac Minis are coming later on this week, apparently, and they will also have M4 CPUs.
458.
by nextos 2y ago
For an alternative POV, quite critical of UMAP, see: https://doi.org/10.1371/journal.pcbi.1011288
459.
by nextos 2y ago
You can see this is a problem if you mine out the distribution of p-values from articles. Andrew Gelman had a great post on this topic I can't find now. Pre-registration could be a great solution. Negative results are also important.
460.
by nextos 2y ago
Mozart/Oz, see the CTM book. Dafny and F* are also evolving pretty nicely.
461.
by nextos 2y ago
Pyro was created by Uber AI Labs. Actually, by Geometric Intelligence, which was eventually acquired by Uber. Geometric Intelligence was founded by Gary Marcus, Zoubin Ghahramani and others. They also had Noah Goodman onboard. AFAIK, Pyro w
462.
by nextos 2y ago
> Could we simply not replace Pyro with Stan for statistical modelling (whether with PyTorch or JAX backend)? Stan has a fantastic NUTS Monte Carlo implementation. Pyro & NumPyro are more focused on variational inference. For a third
463.
by nextos 2y ago
Pyro has a nice normalizing flows tutorial: https://pyro.ai/examples/normalizing_flows_intro.html
464.
by nextos 2y ago
South West France, Spanish Central Pyrenees or North West Spain. Very cheap housing in some select areas, mild weather and great food.
465.
by nextos 2y ago
Living systems are not hyper efficient, because efficiency leads to fragility, and they have evolved to survive on many scenarios.
466.
by nextos 2y ago
> Why is most AI work done in Lisp? Yann LeCun developed Lush, which is a Lisp for neural networks, during the early days of deep architectures. See https://yann.lecun.com/ex/downloads/index.html and https:&#x
467.
by nextos 2y ago
Maemo was indeed fantastic! SFOS is somewhat getting where Harmattan was before Nokia imploded. It's already better in some ways.
468.
by nextos 2y ago
Luca Cardelli has a very famous article elaborating on this topic, which I recommend for a slightly more hopeful perspective: https://worrydream.com/refs/Cardelli_2005_-_Abstract_Machine...
469.
Programming Paradigms for Dummies [pdf](webperso.info.ucl.ac.be)
2 points by nextos 2y ago|0 comments
470.
Simply Logical: Intelligent Reasoning by Example(book.simply-logical.space)
2 points by nextos 2y ago|0 comments
471.
by nextos 2y ago
Do you use Prolog in Academia or you have moved to Industry?
472.
by nextos 2y ago
NixOS is great and IMHO the lowest maintenance effort of any OS if you can stick to well-packaged software. It's also low effort if you need to package simple things yourself. However, if you have to use something slightly unusual, it
473.
by nextos 2y ago
Exactly, good systems do not rely on willpower. They rather make obvious habits effortless. Deviating from the mean is hard. Bad food and sedentarism are the norm.
474.
by nextos 2y ago
BLAST estimates are derived from extreme value theory and large deviations, which is a very elegant area of probability and statistics. That's the key part, I think, being able to estimate how unique each alignment is without having to
475.
by nextos 2y ago
AlphaFold is excellent engineering, but I struggle calling this a breakthrough in science. Take T cell receptor (TCR) proteins, which are produced pseudo-randomly by somatic recombination, yielding an enormous diversity. AlphaFold's pr
476.
by nextos 2y ago
> APL could work well with gpus I've seen at least an APL implementation running on top of Julia, thanks to macros. Julia has good GPU support, and it makes it easy to compose that support with any library. However, kdb+ and q, whic
477.
by nextos 2y ago
Thanks, that's right, 2021 not 2023. Corrected.
478.
by nextos 2y ago
But, as you said, Higgs got his prize once theories were tested. Hence, theoretical contributors (still alive as per prize rules) could have been included here as well.
479.
by nextos 2y ago
I find the prize a bit odd this time since it focused on Hopfield networks and Boltzmann machines. Picking those two architectures in particular seems a bit arbitrary. Besides, Parisi got the prize last year (edit: actually 2021, time flies
480.
by nextos 2y ago
Ruby has evolved slowly language-wise compared to C++, or even Python. Most changes have been in libraries and interpreter / VM implementation. Updating your knowledge from Ruby 1.8 (mid 2000s) to 3.x (current) takes little effort. But

More ›