HNHacker News
TopNewBestAskShowJobs

tgbugs

1,061 karma · joined April 11, 2019

submissionscomments
tgbugs··on We're not going to run out of new anatomy anytime soon
As part of an NIH consortium I work with two teams that are collecting multi-scale anatomical data on the human vagus nerve with one of the objectives being to start to get a handle on the variability between individuals. The variability that the experimental teams are seeing is beyond anything I expected, though admittedly my assumptions were naive. The branching structure and routing of the nerves is basically unique per human and we are in the processes of determining whether there are invariant rules (e.g. for branch ordering) that apply across all individuals. And that is at the level of gross anatomy. So we aren't even done with gross anatomy, despite many biologists thinking that the foundations are complete and have been since the 16th century. Turns out that if you want to be able to apply our knowledge of gross anatomy for more complex clinical use cases we need significantly more data about basic variability in structure so that we know what additional data we need to collect for each individual.
tgbugs··on GPTs and Hallucination
Are you surprised that jpg compression algorithms can reproduce input data that bears striking resemblance to the uncompressed input image across a variety of compression levels?
tgbugs··on GPTs and Hallucination
I think the answer is actually quite clear and rather boring. In order to get something "right" there has to be some external standard of knowledge and correctness. That definition of correctness can only be provided by the observer (user). Alignment between the user's correctness criteria and generated text happens entirely by accident. This can be demonstrated by observing a correlation between coverage of a domain in the training data and the rate at which incorrect results are produced (as discussed in other comments). That is, they get things "right" because there was sufficient training data that contained information that matched the user's definition for correctness. In fact, exceptionally boring.
tgbugs··on Dynamicland 2024
One of the most striking things about this for me is how clearly it demonstrates the fact that community and sharing are not the same thing.

Building tools that enable communities to share effectively seems like another additional challenge, and the fact that virtual spaces and digital spaces are dismissed seems like it might prove a major roadblock to connecting and sharing in a larger inclusive community.

Given the interest in leveraging this for doing science it also seems that this is at risk for empowering individual labs while leaving all interfaces to the rest of the larger scientific community dependent on the current utterly broken system of publication.

tgbugs··on Racket Language
One reason is that it has one of the best cross platform native gui solutions out there. It exists because the research side was focused on teaching programming and needed a solution that their students could just install. There is a bit of a learning curve, but once you're over the hump it is just a pleasure to work with. See also things like https://docs.racket-lang.org/gui-easy/index.html.
tgbugs··on Adobe is buying videos for $3 per minute to build AI model
Amusingly this policy might end up meaning that the AI model they produce by training on this data will never be able to produce video that will ever be worth more than 3$ per minute. They are probably unintentionally filtering for content created by people willing to sell for that price or below, and that bias will be present in any downstream model. You get what you pay for I guess?
tgbugs··on Scientists have traced human tail loss to a short sequence of genetic code
Vitamin synthesis is one of the most pointless things to add because it is something that is lost in many lineages over evolutionary time. The reason is that as long as there are environmental sources of a molecule the pressure to retain the ability to synthesize that molecule natively is zero or even negative, because the synthesis pathways for some of those molecules are quite nasty. You could do it, but those people would probably have increased cancer rates due to running nasty biochemistry in their own cells and eventually their descendants would lose the ability the synthesize it again because vitamin C is naturally present in their environment unless they are e.g. British sailors stuck on a boat for months with a evolutionary strange diet.
tgbugs··on Engineer Used Water Pump to Get $1B Stuxnet Malware into Iranian Nuclear Plant
Depends how many centrifuges they had to destroy to develop it.
tgbugs··on Zeiss's "Holocam" turns glass windows into cameras
I'm super excited about whether something like this could be used to prevent bird strikes on windows. If anyone wants to grow the market for this technology there are likely square miles worth of glass that could have this installed to prevent birds from being killed.
tgbugs··on A Core Calculus for Documents [pdf]
They explicitly exclude computational notebooks in this paper (see page 5). While they don't mention org-mode I'm fairly certain that a significant portion of org functionality can be understood using their concept of reforestation, at least for org-export functionality.
tgbugs··on A Matter of Millimeters: The story of Qantas flight 32
My internal alarm bells started going off as soon as I read about datum AA and datum M. Shouldn't it be possible if not standard practice for the design software to issue a giant warning if you have a part that is defined by two datums that are almost but not quite the same? If they aren't the exact same datum then something like this will inevitably happen.
tgbugs··on Artificial intelligence can revolutionise science
> Our key insight is that the process of citation needs to be handled outside the LLM.

I can imagine taking the citation tree and using the LLM to compact the hypotheses, results, etc. for each node in the tree and sticking that in the vector database could get you pretty far.

tgbugs··on Artificial intelligence can revolutionise science
We are still quite far from being able to implement these kinds of things at scale.

For the literature review piece the key problem is that LLMs are exquisitely bad at working with even the simplest kind of scientific evidence: citations [1, 2]. They will get better, but it is not clear that LLMs can deal effectively with the very sparse kind of evidence that appears in the literature. Also, generating hypotheses isn't exactly the rate limiting step, the bigger issue tends to be when you get people with pet projects/hypotheses in positions of power that dictate funding priorities (e.g. the decades long Alzheimer's Aβ disaster).

For automation and instrumentation of labs the vision is on point and there is interest, and active work, if not large amounts of funding, to bring that vision to reality [3, 4, 5]. However, we simply don't have the tooling needed to be able to express the full complexity of experimental protocols in a way that can be verified. Sure you can write a python script to control a robot, but it is exceptionally difficult to extract the scientific meaning from that.

My PhD work was to develop a formal language for scientific protocols, and I'll be continuing to develop it, but there is still a long way to go.

1. https://doi.org/10.7759/cureus.39238 2. https://doi.org/10.1016/j.mcpdig.2023.05.004 3. https://www.youtube.com/watch?v=_gXiVOmaVSo&t=865s 4. https://doi.org/10.1109/JIOT.2020.2995323 5. https://ccc.ucsf.edu/sites/ccc.ucsf.edu/files/Marshall_W_CCC...

tgbugs··on Why does the USA use 110V and UK use 230-240V? (2014)
The tradition answer to this question is surely this wonderful Tom Scott video: https://www.youtube.com/watch?v=UEfP1OKKz_Q.
tgbugs··on How was the size of Earth first measured? (2015)
Ah what a fantastic presentation. It answers a lingering question about the use of spectroscopy and apparent brightness to determine the distance to a star. The key step that I had missed was using the stars with distances measurable by parallax to calibrate the color/apparent brightness so that it could be used for stars that were beyond the distance for parallax. Obvious in hindsight, but wonderfully and clearly explained here.
tgbugs··on Ask HN: Is anyone using PyPy for real work?
I don't have anything rigorous, but I can say that I see the usual ~4x speedup when using rdflib to parse large files so that a 20 minute workload in cpython drops to 4 or 5 minutes when run on pypy3.

I just reran one of my usual benchmarks and I see 2mins for pypy3 (pypy 7.3.12 python 3.10.12) peak memory usage about 8gigs, 4.8mins for python3.11 (3.11.4) peak memory usage about 3.6gigs (2.4x speedup). On another computer running the exact same workload I see 6.3mins and 19mins (3x speedup) with the same peak memory usage.

I don't have any numbers on the dataset pipelines because I never ran them in production on cpython and went straight to pypy3. It is easy for me to switch between the two implementations in this context so I could run a side by side comparison (with the usual caveat that it would be completely non-rigorous).

I also have some internal notes related to a project that I didn't list because it isn't public, isn't in production, and the benchmarks were collected quite a while ago, but I see a 4x increase in throughput when pulling large amounts of data from a postgresql database from 20mbps on cpython 3.6 to 80mbps on pypy3.

tgbugs··on Ask HN: Is anyone using PyPy for real work?
We use pypy3 on musl via gentoo in production to run dataset validation pipelines. The easiest place to see that we use pypy3 is probably [1]. The build process and patches we carry are under [2].

We also use pypy3 to accelerate rdflib parsing and serialization of various RDF formats. See for example [3].

Thanks to you and the whole PyPy team!

1. https://github.com/tgbugs/dockerfiles/blob/6f4ad5d873b7ab267...

2. https://github.com/tgbugs/dockerfiles/blob/6f4ad5d873b7ab267...

3. https://github.com/SciCrunch/sparc-curation/blob/0fdf393e26f...

tgbugs··on It's 2023, so of course I'm learning Common Lisp
For static builds, if you're willing to run a slightly older version of sbcl daewok's work on building and linking sbcl in a musl environment might be solution you're looking for. I've tried to port his patches to more recent versions but there are segfaults due to changes in upstream.

https://www.timmons.dev/posts/static-executables-with-sbcl.h... https://www.timmons.dev/posts/static-executables-with-sbcl-v...

tgbugs··on An invariant from category theory solves a problem in mathematical ecology [pdf]
Wow. The fact that there is an objective answer that is independent of any perspective on the importance of rare species is a rare gift, at least for this part of the problem.

Some questions and thoughts.

It seems that the result could vary based on how you construct the similarity matrix Z, e.g. is it purely taxonomic? or does it try to account for the ecological roles that a species is playing in the community, etc.

A seeming limitation is that the optimization works only for a fixed set of n species. While it is useful for managing existing communities, it means that there is still a question of whether larger n is strictly better, and leaves open questions of how to deal with transient or migratory members (if the community is spatially bound).

The answer I think, is that it depends on how the similarity matrix is constructed. If every species is fully dissimilar then increasing n is always a good thing. If you use niche space to construct it and new species do not some add or enter new niches so they overlap with others, then they will be close to another species in the matrix and increasing n will not have much impact. On the other hand if you use a purely taxonomic approach then you wind up balancing the number of birds and mammals regardless of niche.

It is not clear to me whether it is possible to construct a similarity matrix that can account for the interaction between n, the carrying capacity of the ecosystem, and the number of available niches (or the ability of species to create new niches). By analogy if you have a stream (sunlight) powering water wheels, how many wheels and how many levels of gears (layers in the ecosystem) can be added, created, and/or sustained? At what point does adding an additional species mean that either two species are forced to be close together in the similarity matrix or both their populations must shrink in size because they must compete for the same energy sources?

Does the model sometimes produce impractical results, e.g. that it is good to have a single member of a sexually reproducing species (this is probably an orthogonal concern and you would want to scale to real population sizes such that the minimum corresponded to the smallest viable a self sustaining population)?

Is there evidence that maximizing diversity using this measure actually produces more robust and stable ecologies?

tgbugs··on In practice, cool URLs can become inaccessible even if they don't change
When I was thinking about this problem a couple of years ago (eek, time flies!) I came to the conclusion that the only archival approach that was viable for this was to have a dedicated url resolver service that was independent of the DNS system and could be swapped out.

Obviously you wind up with resolver resolver resolver services if this goes on for too long, but it is one of the few workarounds for the fact that time (and many other things) are not reified in the DNS system.

tgbugs··on Racket frustrates me
On Gentoo xgqt pretty much has everything working as a portage overlay.

https://gitlab.com/gentoo-racket/gentoo-racket-overlay

tgbugs··on Racket: Lisp for the modern day
Racket has had significant engineering and design time spent directly on the use case you describe. See this old thread from John Carmack about his son learning with Racket https://news.ycombinator.com/item?id=10111479. Things have continued to improve since then as well.

From my own experience, it is absolutely possible to get good interaction with low latency in Racket, but of course as with any language you have to have enough experience to know how to avoid performance pitfalls.

For some of the gui work I've done I needed to mix threads and semaphores to get things to be really responsive when there could be a long running process triggered by a UI interaction, but in Racket it has been vastly easier and safer than in other languages.

tgbugs··on Racket: Lisp for the modern day
Looks like you're already in Emacs. I strongly recommend racket-mode as mentioned in another thread.

With regard to prototyping GUI's I'd suggest taking a look at https://github.com/mfelleisen/7GUI. https://github.com/Bogdanp/racket-gui-easy could also be a good place to start.

With regard to Racket more generally, I'm probably not the best person to ask since I had a very high friction start where I just banged my head against the wall until things made sense.

tgbugs··on Racket: Lisp for the modern day
I've been using Racket for about 8 years now, and it is now my goto language for anything that needs a gui. It took writing 3 or 4 not-so-little guis to feel comfortable working with Racket's object system, but now it basically goes at the speed of thought.

Lately I've been using Racket to prototype the interaction design for a game, and I had written some placeholder text. I had put fake interpolation in for {player-name} and things like that, and then I realized that #lang at-exp would let me just ... do that, as described in the original article.

I had to laugh, because I maintain a bunch of Python code, and if Python has batteries included, Racket comes with a retrofittable fusion battery pack.

One other little anecdote, is that I've been porting an old plt-scheme codebase to Racket CS. In the second phase I have started replacing an ancient C++ rendering layer with the racket gui canvas functionality. During the process I moved hundreds of runtime type checks in the C/C++ layer to the Racket contract system. It was as if the Racket team had somehow secretly obtained access to the codebase I was working on and had designed the perfect abstractions to solve the problem.

tgbugs··on The Embeddable Common Lisp (2005) [pdf]
If you are willing to use somewhat older version of SBCL (2.1.11) then you can use daewok's static executable branch and link against musl. That version only supports gzipped images, but even so you can get it down under 15mb for small programs.
tgbugs··on Ancient lead pollution in a Roman harbor (2017)
Lest Darkness Fall by L. Sprague de Camp would be a classic answer (though it is also scifi so maybe not quite what you are asking for).
tgbugs··on Directly compiling Scheme to WebAssembly: lambdas, recursion, iteration
It will depend on which implementation specific behavior a given library uses. For example, guile has `#{}#` syntax for verbatim atoms. If a library uses that it will not work on any other scheme. Guile and gambit also have divergent behavior when e.g. parsing char literals `#\)a` parses correctly in guile but not gambit, and on the other hand `'a'b'c` reads as `c` in gambit but `a'b'c` in guile. And these are just issues with variant behavior in the reader.

The list of these kinds of edge cases is long so it probably winds up being easier to try to just run a library to see if it works. I doubt anyone has done a systematic survey.

tgbugs··on PyPI Was Subpoenaed
One theory that I don't see mentioned yet is that someone used an upload to pypi to exfiltrate data or simply as a way to upload arbitrary data somewhere. In a sense pypi is just a file hosting service, so it could have nothing to do with any actual python projects at all.
tgbugs··on Ask HN: What's your favorite software testing framework and why?
Racket's (module+ test ...) by far.
tgbugs··on Htmx Is the Future
I first encountered the principles behind htmx in its precursor intercooler.js. Those principles really resonated with my distaste for complexity. Amusingly I found out about htmx itself when rereading https://grugbrain.dev and it all clicked! htmx is crystal that trap internet complexity demon!
← PreviousPage 2 of 9Next →