pama

43 ms·

811.
by pama 3y ago
Are there any startups in the space of probabilistic hardware for ML? How hard would it be to make something minimally useful and cheap even if it wouldn’t compete with nvidia for most use cases?
812.
by pama 3y ago
It’s not the programming. Every default in matplotlib is poor. You have to write about ten times more lines of code to get something clean with legible fonts and minimal, understandable aesthetics. Very few people bother and then you get p
813.
by pama 3y ago
To some degree yes, but these particular rehydrated antibody tests leave very little room for genuine false positives. If something binds on the test part of the strip it has the right epitope and there are no other viruses yet with these
814.
by pama 3y ago
Apologies for the confusion. Yes, run the gel analysis using Fiji. (Fiji is is a package distribution of ImageJ.)
815.
by pama 3y ago
If anyone cares to increase the resolution of a rapid test, I recommend using Fiji and treating the image of the test strip like an antibody blot image and run the quantitative gel analysis (easy to find info on the web). The software integ
816.
by pama 3y ago
Please read more carefully. OpenAI never “declared they trained on pirated works.”
817.
by pama 3y ago
Totally minor but this website does not show anything in lockdown mode on the iPhone.
818.
by pama 3y ago
Most of the air pollutants are generated inland, and most of the poor quality air remains somewhat local (or within a couple hundred miles) from the source. Oceans provide a lot of space for the particles to diffuse and eventually get captu
819.
by pama 3y ago
They mention replication and optimistic concurrency to allow multiple writers to work in parallel.
820.
by pama 3y ago
Not sure exactly what you mean here but most of the quantum mechanics calculations in material science and chemistry applications can now be done at dramatic speedups using such approximate models. It has been exactly this Bloom Filter mome
821.
by pama 3y ago
Why didn’t you ask the authors for the actual sample dataset? Wouldn’t the commit to delete these files indicate that they were not the ones used in the study?
822.
by pama 3y ago
Isn’t that a different thread though?
823.
by pama 3y ago
I quit Reddit today in part because of this event and I had never used Apollo. I started reading it more regularly 5 years ago because a friend had ranked in their top twenty and got me curious about what he spent his time on. I loved obser
824.
Wildfire smoke live updates from the NYTimes(nytimes.com)
5 points by pama 3y ago|0 comments
825.
by pama 3y ago
Read, think, and sleep. Problems in this category often need time in your brain to settle and clarify. The process is rewarding, and after you’ve tackled a few such problems (either proving them impossible or actually solving them) you gain
826.
by pama 3y ago
My main argument would be size of the good models and bandwidth, not speed of the computation. As long as it’s hard to transfer 1TB to a remote location we are still moderately safe as a society and a large model of the scope of PB can tak
827.
by pama 3y ago
Thanks! I think this future will be beautiful and hopefully close to achieving the dream of groups of AI helping each other to improve themselves much faster than humans could improve them. The mental model that rogue AI evokes in me is th
828.
by pama 3y ago
I wish GPT-4 could make functional copies of itself on my computers and I’d find ways to allocate clusters for it to do so if it were possible, but until we see that level of development I don’t think that autoGPT and variations are a serio
829.
by pama 3y ago
Proudbarrassed.
830.
by pama 3y ago
Unfortunately even though the questions are about GPT-4 the answers and personal experience only refer to GPT-3.5 at most. I hope openai changes the name of the next version to avoid this confusing narrative by prominent people. 3.5 vs 4
831.
by pama 3y ago
Table 2 of the OpenAI report had 7 public benchmarks and figure 5 had another 27.
832.
by pama 3y ago
I have not answered unknown call numbers by using my voice for about ten years now. I start muted or I whistle to see if a spammer is on the other end, but mostly I simply let them go to voicemail, then block and report the caller. It was c
833.
by pama 3y ago
The model does not keep training every day on the current data. It would be nice if it could but no sign this actually happens. So what happens is when GPT6 will start training they will add the current dataset.
834.
by pama 3y ago
Thanks!
835.
by pama 3y ago
Separate question: what is the eventual monetization strategy? Of course, feel free to not answer, though if that plan is not a secret and makes sense to people, it might help spread the technology.
836.
by pama 3y ago
Are there plans to port any major existing Python libraries to pure mojo (mojo that will not use CPython)? Is there a plan for a general purpose matrix or tensor library like numpy or pytorch, but in pure mojo?
837.
by pama 3y ago
Congrats on the launch! In addition to the dramatic possibilities for ML, I hope mojo will have an impact in other scientific software through the ease of specialization to hardware capabilities. The current common subset of Python and “pu
838.
by pama 4y ago
Present. I think there exist many of us. Chemistry is a very wide field though, so not sure if organic synthesis vs theoretical chemistry vs physical chemistry vs biochemistry will end up more useful to help tackle drug discovery problems
839.
by pama 4y ago
It is not the determinism that’s a concern. JavaScript can test the capabilities of your machine and then create a new connection to transmit its results back to another computer. Even in the hypothetical scenario when you are only allowed
840.
by pama 4y ago
You are welcome.

More ›