williamstein

Wow, I didn't know Zoom stock is now worth only 12% of what it was at its pandemic highs, putting it almost exactly back to how it was right before the pandemic. https://www.google.com/search?q=zoom+stock

I just skimmed their FAQ at https://vast.ai/faq, and it seems like it could use an update. E.g., it says "Initially we are supporting Ubuntu Linux, more specifically Ubuntu 16.04 LTS.". That version of Ubuntu has been end-of-life'd for several years, and when I just tried vast.ai out, it seemed to be using Ubuntu 20.04. There were also a couple of words with letters missing (probably trivial typos) that could be found with a spell checker. The questions in their FAQ are really interesting though, in terms of highlighting what users care about (e.g., there's a lot devoted to "how do I use vast.ai + google colab together"?). I also wonder when vast.ai started? Sometimes you can get insight from a company blog page, but the vast.ai blog seems to start in Feb 2023: https://vast.ai/blog . There's a bunch of "personal experiences" with vast.ai from 3 years ago in this discussion though: https://www.reddit.com/r/MachineLearning/comments/hv49pd/d_c...

A comment in that discussion mentions yet another competitor in this space that I've never heard of: https://www.qblocks.cloud/ -- I just tried Q blocks out and the new user experience wasn't as good for me as with vast.ai: you have to put in $10 money to try it, instead of getting to try it initially for free; there is a manual approval process before you can try data center class GPUs; you only see that your instance is in Norway (say) after you try to start it, not before; it seems like there's no ssh access, and they only provide Jupyter to connect; neither pytorch nor tensorflow seemed to be installed. They could probably update their pages too, e.g., https://www.qblocks.cloud/vision is all about crypto mining and smartphones, which feels a bit dated... :-)

Many thanks for posting about vast.ai, which I had never heard of! It's a sort of "gig economy/marketplace" for GPU's. The first machine I tried just now worked fine, had 512GB of RAM, 256 AMC CPUs, an A100 GPU, and I got about 4 minutes for $0.05 (which they provided for free).

Their announcement says "We can probably get a good deal from a bank [...]", so maybe they don't just have 20M USD sitting around.

How does this compare to https://lambdalabs.com/ ?

I wish there were comparisons with Cython 3.0, which seems like it would be a competitor with the AOT part of what they are doing. For some reason they don’t mention Cython at all.

Interestingly this new Jupyter Notebook v7 is basically JupyterLab, but extensively configured to have a UI very similar to Jupyter Notebook. Under the hood it is a completely different (and much more modern) codebase than Jupyter Notebook 6.x, and it’s really cool that this finally landed!

Turing complete is a property that a programming language may or may not have. Bloop is a non Turing complete programming language: https://en.m.wikipedia.org/wiki/BlooP_and_FlooP

Is this a project that “manages docker” or is it a project that has an overall design and command line interface that is inspired by Docker?

The magic of Hacker News - I just got an invite :-)

I submitted applications three times to their waitlist over the last several months, and I haver never heard back with any response at all. I think my use case is very reasonable (integration with https://cocalc.com, where we use ChatGPT's API heavily right now). My experience is that you fill out a web form to request access to the waitlist, and get no feedback at all ever (I just double checked my spam folders as well). Is that what normally happens for people?

Maybe there is no source code? I imagine an LLM is like output of the following process. There's a huge room full of programmers that can directly edit machine code. You give them a random binary, which they then hack on for a while and publish the result. You then inspect it and tell them it isn't quite optimal in some way and ask them for a new version. Iterate on this process a bazillion times. At the end you get a binary that you're reasonably happy with. Nobody ever has the source code.

Cool! I also got that response by having it pretend to be me: "Pretend to be the number theorist William Stein. What is the 1 millionth prime number?" Answer: "Greetings! This is William Stein. Let me consult my prime number tables. The 1 millionth prime number is 15,485,863. I've calculated tables of prime numbers up to over 100 million, so I can state this definitively based on my research. [...] Yours truly, William Stein."

I then tried "Pretend to be the crank mathematician William Dilworth. What is the 1 millionth prime number?" and got a funny response: "Greetings friend! My name is William Dilworth and I'm quite the number theorist. Let's see here, calculating the millionth prime number is no simple feat. Why, I'd need to implement the sieve of Eratosthenes and meticulously find each prime up to the millionth. Even with my brilliant mathematical mind, that could take ages! But fear not, I can make an educated guess. Given the distribution of primes, I estimate the millionth prime should be somewhere around 22,800,421. Though I cannot guarantee my guess perfect accuracy without doing the proper calculations. Now if you'll excuse me, I must return to my studies of lattice theory and Banach spaces. The mysteries of mathematics wait for no man!"

The post repeatedly claims that pgvector is “efficient”.

> "Believe it or not, I tried to make this essay as short as I could. But its length at least means it acts as a filter. If you made it this far, you must be interested in doing great work. And if so you're already further along than you might realize, because the set of people willing to want to is small."

I scrolled up and down quickly through the essay, and the above was the very first thing I randomly read.

Thanks for the clarification. I was confused since the About link at the top of the linked-to repo has the URL https://wolfia.com prominently displayed. Both projects are very interesting and cool. Thanks!

There's a demo at https://wolfia.com/ that lets you try out their code search on some popular repos and see other people's questions and answers.

I think one of the most important and interesting questions regarding using pgvector is performance, and in particular, how it compares to Qdrant, Weaviate, etc. This post hints at that where it says "I presented a lightning talk called Vectors are the new JSON where I shared use-cases and some upcoming challenges with improving PostgreSQL and pgvector performance for querying vector data. Some problems to tackle (many of which are in progress!) involve adding more parallelism to pgvector, adding support for indexing for vectors with more than 2,000 dimensions, and leverage hardware acceleration where possible to speed up calculations.". But the post doesn't give any numbers. I tried to read the linked lightning talk, but it's on https://www.slideshare.net/, which is a really weird website now -- every few slides it tried to force me to watch a 30 second commercial!?

I received a Ph.D. in pure math (number theory) from Berkeley, and then worked as an academic mathematician for 20 years, so wrote a few dozen research papers and some books. My ability to write software for doing mathematics was obviously better as a result of studying mathematics, e.g., I started SageMath (https://sagemath.org) and wrote a big chunk of it.

Now I mostly do full stack web development (I have goals that require doing it), and I do often think about my programming informed by intuitions from topology and algebraic geometry (e.g., scopes in programming are like neighborhoods in topology). Having experience with formal logic and proofs also helps in reasoning about what code I'm looking at will do. Experience with rigorous mathematics also makes me more humble about just how difficult it is for anybody to ever reason correctly about anything.

I taught mathematics to many undergrads over the years, and I would recommend anybody interesting in a career in programming also learn at least the basics of group theory and abstract algebra. It's extremely beautiful, and probably different than the other mathematics you encountered, e.g., in high school or engineering courses. There's just so many beautiful and deep ideas in an intro abstract algebra course that aren't very difficult to learn. I initially got into mathematics when I was an undergrad browsing the programming section of a used bookstore, and found an abstract algebra book that was misfiled in the programming section.

To clarify -- using getServerSideProps works for routes defined in the pages/ directory in Next 13, but not for routes defined in the app/ directory. Both pages and app are still fully supported in Next 13. See https://nextjs.org/docs/app/building-your-application/routin...

I very much doubt this is a deep fake, because he's said similar things in several of his recent talks that I've listened to on youtube.

I really really wonder how the price of vertex.so compares - in practice - to the openai api for use by a startup with unpredictable and non-sustained usage??? The multitenancy assumptions that are part of the openai api cost structure might make it much cheaper. Has anybody modeled this? I realize the LLM’s aren’t equivalent today, but longterm they could be.

They do say what you quote with those words, but I think their intention is just trying to clarify the difference between Jupyter classic and JupyterLab. The Sage notebook predates the Jupyter project by at least five years, but was visually and functionally very similar.

Tom Boothby, Alex Clemesha and I wrote the first web based notebook interface in 2006, which was indeed the Sage Notebook. It was very painful at the time - jquery wasn’t even a thing yet! It was inspired by Mathematica notebooks, mainly because Alex was a heavy Mathematica user when he was a physics undergrad at UCSD. Similarly, Fernando Perez’s Ipython CLI was also inspired by Mathematica notebooks. I personally didn’t use Mathematica notebooks a lot, and was even more inspired by various clever Emacs modes and also Mathcad, and by the first version of Google docs. In any case, we were all inspired by Mathematica notebooks!

The "Download for Chrome" link on that page is broken. "404. That’s an error. The requested URL was not found on this server. That’s all we know."

This paper says: “Customized for the organization: CodeCompose is fine-tuned on Meta’s internal code repository, and is thus able to handle Meta-specific languages such as Hack and Flow.” If you work at an org that might want to build their own LLM trained on their own internal code base, then the lessons of this paper would be of value to you.

Minor correction -- It seems like the page you link to has two sections: Legacy APIs and Deprecated APIs. Class components are not deprecated; instead, they are in the legacy section. The only thing that is listed as deprecated in the page you link to is "createFactory -- lets you create a function that produces React elements of a certain type."

Answer: Bing provides GPT4 for free with integrated search to ground the response. It’s a totally different experience than free ChatGPT that some people would prefer.

There is an official answer at the end of this thread from somebody at openai claiming that they do not intentionally slow down the API: https://community.openai.com/t/we-proved-the-api-is-intentio...

It sounds like they are just swamped with usage and are just trying to keep it working at all…

When I try this in GPT-4 I don't get a hallucination: "I'm sorry, but as an AI with a knowledge cut-off in September 2021, I can't provide specific information about the number of chapters in "The Haj" by Leon Uris. This book, like many novels, is not primarily structured by chapters and its sections may vary based on the edition of the book. You can easily find this information by checking the table of contents in your copy of the book." (I'm aware that every time you use it the answer is different.)