btrettel

Location: United States (Open to any US location)

Remote: Yes, open to remote, hybrid, or in office

Willing to relocate: Yes

Technologies: Fortran, Python (including Matplotlib, Numpy, Scipy), OpenMP, Git/GitHub, Linux, Bash, others...

Résumé/CV: Available on request

Email: http://trettel.us/hn-jobs-email.png (to reduce spam, please type in the email address)

GitHub: https://github.com/btrettel

Personal website: http://trettel.us/

I'm Ben Trettel, a mechanical engineer with about 15 years of experience, specializing in computational fluid dynamics, multi-disciplinary design optimization, and verification and validation of computer simulations.

I am particularly interested in opportunities to build cutting-edge physical products where computational simulation and design optimization are key.

Please email me at the address above about only specific job opportunities that are matched to my skills and background. Do not email me about generic job opportunities that anyone can apply to or non-job opportunities like advertising your job board or giving me advice on my resume.

I am not interested in gig economy type contract work.

Recruiters: Send full details of the position including company and job description in the first email. As long as the position meets the criteria above, do not ask if I am interested, just send the details.

It might be worth putting "(2013)" in the title. At first, I thought the title was clickbait about using LLMs without thinking, but it's from well before LLMs and makes an interesting point. I wouldn't say that "Stop Thinking" is the right way to phrase the author's point, though, but it does get attention.

I'm the administrator of 3 online forums, 2 of which are basically archives at this point. All 3 forums were getting absolutely hammered by bots. In fact, all 3 were taken offline by their hosts due to the heavy traffic in 2026. I installed Anubis on 2 of them and bot traffic decreased dramatically to the point where it appears to be only legitimate bots. (Unfortunately, the other forum couldn't use Anubis so easily and has remained offline since then.)

It's clear to me that Anubis won't work forever. You don't need to get as sophisticated as this blog post either as you can bypass the proof-of-work challenge with the right user agent string. I think the reason I'm not seeing bots change their behavior in response to me adding Anubis is simply that my forums are relatively small players. To the people running these scraper bots, a handful of forums adding Anubis is no big deal to them as there are tons of other sites they can scrape.

I think turning the forums that are archives into static HTML is the best path forward for those forums, but that's easier said than done. I've also been thinking about writing a new forum software where all of the content pages are static HTML, and the only dynamic pages are where dynamic features are unavoidable like when posting.

I visited yesterday according to my browser history. I contributed last probably over a year ago, but I don't know off the top of my head.

The Library of Congress did in fact switch their catalog to a new system in the past year or so, and I think some breakage is expected. It would be more productive to report this issue to them than complaining in a blog post they probably won't notice. They've been receptive to feedback in my experience. I reported a cataloging issue to them before and the issue was fixed. There's a link titled "Feedback" at the bottom of every catalog page.

Also, I'm not sure why the author is maintaining their own bibliographic database directly in Sqlite. Something like Zotero would probably be a better fit for them. (I'm aware Zotero uses Sqlite, that's not the point.) Zotero's bibliographic data importer works fine for the Library of Congress, even now.

Edit: I reported the bug to the Library of Congress myself:

> In the catalog, for seemingly every record, if "MARC record" is selected, "Download" is pressed, and any of "Dublin Core", "MODS", "MARC XML", or "Bibframe" are selected, a 404 error is returned. Those forms of cataloging records can not be downloaded and should be fixed. Thank you.

Edit #2: I got a reply within 5 minutes:

> Thank you for taking to the time to report this issue. We have noted the problem and are looking into it and hope to have a fix in a future release.

Edit #3: And rather than reply to a blog post in a place the author won't see, I've emailed the author about how he can report issues to the Library of Congress, and noted that I've already reported this bug.

You could consider strategies used by engineers in high-income countries when faced with outsourcing/offshoring/immigration:

- Work in industries that require a local human, licensing, and/or citizenship. Law, defense, civil engineering, etc. It's not clear to me how robust these requirements are in the future, however.

- Set up legal barriers to prevent people from using AI to replace you. (The equivalent mostly didn't work for the engineers in high-income countries in my view. Note that I am not advocating for this, just listing it as an option.)

- Start your own business where you aren't directly doing engineering projects in return for money. Could be selling products direct to consumers instead, for example. Engineering can still be involved, of course. (Most likely to be successful in my view.)

I have encountered the same problem myself. I've switched to linking to an image of a new disposable email address. Zero spam since then. (Actually, zero non-test messages total since then, but that's fine too.) I suppose that will work only as long as not too many people do it, though.

However, there are legitimate companies hiring through those threads. I know someone who got their current job from one of those threads. If you can avoid the spam, I think the benefits outweigh the downsides.

Your experience is not universal.

I used modern Fortran in undergrad and grad research from 2011 to about 2015. I rarely ever encountered FORTRAN 77 back then, and never had to write any. I worked as an intern at NIST and LANL during that time too and it was modern Fortran there as well. At NIST, as far as I'm aware, that group had been using modern Fortran for probably a decade or more before I started in 2011.

I didn't have to write any FORTRAN 77 until 2025 in my current job when I was given a legacy code to work with.

Scammers seem to use Hacker News heavily. Presumably the moderators keep the scam messages off Hacker News itself, but the moderators can't stop scammers from emailing people.

Posting on the "Ask HN: Who wants to be hired?" threads returns a ton of scams, particularly fake recruitment scams. Lightly obfuscating my email address like user at domain dot com didn't help much. Since changing my email address on there to a link to an image, I haven't got any scam messages or spam (yet).

Location: United States (Open to any US location)

Remote: Yes, open to remote, hybrid, or in office

Willing to relocate: Yes

Technologies: Fortran, Python (including Matplotlib, Numpy, Scipy), OpenMP, Git/GitHub, Linux, Bash, others...

Résumé/CV: Available on request

Email: http://trettel.us/hn-jobs-email.png (to reduce spam, please type in the email address)

GitHub: https://github.com/btrettel

Personal website: http://trettel.us/

I'm Ben Trettel, a mechanical engineer with about 15 years of experience, specializing in computational fluid dynamics, multi-disciplinary design optimization, and verification and validation of computer simulations.

I am particularly interested in opportunities to build cutting-edge physical products where computational simulation and design optimization are key.

Please email me at the address above about only specific job opportunities that are matched to my skills and background. Do not email me about generic job opportunities that anyone can apply to or non-job opportunities like advertising your job board or giving me advice on my resume.

I am not interested in gig economy type contract work.

Recruiters: Send full details of the position including company and job description in the first email. As long as it meets the criteria above, do not ask if I am interested, just send the details.

The about page [1] suggests that abstracts are added to the dblp data.

[1] https://pldb.kirancodes.me/about

I added this to my list of specialized bibliographies:

https://github.com/btrettel/specialized-bibs

In my work (scientific or otherwise), I try to avoid dependencies if possible. That's not always possible, so a solution like Docker or NixOS is needed, but the problem can be improved a lot without a technical solution. Either feels like fighting an uphill battle though as most researchers think short term and just pick whatever is convenient in the moment.

The title reminds me of this, which is arguing the opposite direction: https://softpanorama.org/Articles/oss_as_academic_research.s...

This analogy seems strained to me.

The credulous LLM users Dan Luu is complaining about don't take the equivalent of thousands of photos of the same sunset. They take a couple photos at most, say "LGTM", and move on to taking a photo of something else, even if their couple of photos of the sunset aren't great.

As someone who will often compare different things, I think doing that usually takes time. I think LLMs can help reduce the time (especially if it's programming, not necessarily so much in other areas), but I'm not seeing many people use LLMs to try a large number of independent approaches and compare them.

I've thought exactly that "writing is thinking" before as a reason to not let a LLM write for me.

Then again, I've seen a counterargument [1] by someone who clearly heavily uses LLMs for writing (going by both their LLMy writing style and their own admission). The person I'm citing describes a process where they get a LLM to write something, they check over it and provide feedback to the LLM, the LLM rewrites, and the process repeats iteratively. So clearly he is putting thought into the process.

I think there is something valuable missing, even if it's hard to clearly express. I'll try. The threshold for what I'm willing to accept if I'm simply approving something is likely different from what I'll get if I write something myself, for instance. Saying "LGTM" is too tempting. It seems to me like he's outsourcing his selection of topics to cover as well. If you're not thinking yourself about what to cover then it would be very easy to miss a critical subject. There also an asymmetry between checking and generating something with constraints placed on it. Checks can't catch everything, and a constrained generating process can reduce the amount that needs to be checked, avoid issues that can't be checked so easily, and focus your attention on areas that you know historically have had issues with this generating process. I've thought about this quite a bit in terms of whether to write new code or use an existing library. Sometimes "the devil you know" (my code) is better than an existing library simply because I understand its flaws better.

[1] https://www.nature.com/articles/d44148-026-00236-3

I assume this was supposed to link here: https://mathstodon.xyz/@tao/117207849921390904

This was interesting as I had wondered whether Terence Tao realized that the Navier-Stokes Millenium Prize problem was kinda useless from a fluid dynamics application perspective. Turns out that he does!

> While the equations do come from a very natural physical motivation - the study of incompressible fluids - the regularity problem is not important for its direct physical application. [...]

I disagree that the Millenium Prize problem would tend to "spur further development of the field". I'm not seeing that. The people working on the Millenium Prize problem aren't helping solve more practical problems like making better turbulence models. I wish we had more people of Tao's caliber working directly on turbulence modeling! Indirect work won't cut it here. In my view, the Millenium Prize problem has mostly been a distraction from more important issues.

Location: United States (Open to any US location)

Remote: Yes, open to remote, hybrid, or in office

Willing to relocate: Yes

Technologies: Fortran, Python (Matplotlib, Numpy, Pandas, Scipy), OpenMP, Git/GitHub, Linux, Bash, others...

Résumé/CV: Available on request

Email: d6q730rl [at] fastmail [dot] com

GitHub: https://github.com/btrettel

Personal website: http://trettel.us/

I'm Ben Trettel, an experienced mechanical engineer with a PhD, specializing in computational fluid dynamics, design optimization, and verification & validation of computer simulations.

I am particularly interested in opportunities to build cutting-edge physical products where computational simulation and design optimization are key.

Please email me at the address above about only specific work opportunities that are matched to my skills and background, not generic work opportunities that anyone can apply to or non-work opportunities like advertising your job board.

I haven't seen anything like this before, but I think it's a good idea. I would suggest starting it yourself and putting it on GitHub something similar so that anyone can submit more via pull requests.

I have a PhD in mechanical engineering.

I would back up and instead ask whether getting a PhD is a good idea. Unfortunately, academia is a minefield. PhD students are largely cheap labor. Getting a PhD can be a valuable apprenticeship, but often it's abusive and poorly paid. You might nominally get some freedom, but the grant funding wants you to do something you might not care for. The opportunity cost of a PhD is huge. You should consider as an alternative taking a more conventional job, maybe a part time one, and doing science on the side to figure out how to start your business. A part-time engineer making $50K/year is getting a much better deal than the vast majority of PhD students.

The problem-first vs. science-first framing sounds good on paper, but it would be difficult to accurately determine whether a particular lab has either focus ahead of time. What you see from the outside is mostly marketing. Most labs are really neither and instead are publication-first. Publications are the currency valued in academia.

Location: United States (Open to any US location)

Remote: Yes, open to remote, hybrid, or in office

Willing to relocate: Yes

Technologies: Fortran, Python (Matplotlib, Numpy, Pandas, Scipy), OpenMP, Git/GitHub, Linux, Bash, others...

Résumé/CV: Available on request

Email: d6q730rl [at] fastmail [dot] com

GitHub: https://github.com/btrettel

Personal website: http://trettel.us/

I'm Ben Trettel, an experienced mechanical engineer with a PhD, specializing in computational fluid dynamics, design optimization, and verification & validation of computer simulations.

I am particularly interested in opportunities to build cutting-edge physical products where computational simulation and design optimization are key.

Some comments not addressing your question:

I think you are holding climate science to a far higher standard than the other physical sciences.

In the physical sciences, blinded experiments are rarely done. The need for blinding is greatly reduced compared against the medical and social sciences as the data measured is far more objective and observers usually can't influence the results. I've done experiments where I basically start recording data, turn a valve, and from that point on, the result is outside of my control as long as I'm watching from a distance of 10 feet or so. That's often not the case in the medical and social sciences. In my experience, blind experiments in the physical sciences take the form of a blind prediction challenge where a bunch of teams are asked to predict what a certain experiment will do before the experiment is run. This is a good practice that I advocate. I also am interested in blind data analysis and "fake-data simulation" as Andrew Gelman calls it.

There also are a lot of times in the physical sciences where duplicating the "full scale" case is not possible for various reasons (cost, legality for nuclear weapons testing, etc.). Instead they do experiments on scale models (which might not have similitude [0]) or parts of the full problem. This is not ideal, and I do think the researchers could do better, but the situation is unavoidable. Again, there's no reason to single out climate science on this.

I also want to strongly push back on the "computers models are basically mathematical assumptions, not physical laws" part. I'm a mechanical engineer who works in computational fluid dynamics. The models I use have a lot of overlap with climate models, but don't get as much scrutiny even when they are of similar reliability. A large computer model like a climate model has a lot of components, some of which could be regarded as very reliable (likely what you mean as "physical laws") and some of which are less reliable. But even the less reliable components are not assumptions. They always are backed up by some data, perhaps not as a comprehensive as is wanted, yes, but some data.

[0] https://en.wikipedia.org/wiki/Similitude

Drag-and-drop of the PDF file worked for me in Zotero 9 (latest version). I never used this feature before. This would be greatly preferred to my earlier suggestion to get a LLM to generate a list of DOIs.

Zotero is not just a front-end to BibTeX. Here are some things I like about Zotero off the top of my head:

Zotero integrates well with various online services. This I think is Zotero's most valuable feature. The simple fact is that not everything exports BibTeX. Can you get a library catalog to export BibTeX? Perhaps in CS, most publications provide good BibTeX, but a lot of journals I (a mechanical engineer) deal with don't provide BibTeX at all to my knowledge. But I can simply press a button in Firefox and import the bibliographic data embedded in the web page into Zotero. (No, Google Scholar is not a good solution here because it's frequently inaccurate and incomplete.)

I'm a technical guy who can handle BibTeX fine, but I still prefer the Zotero UI over using BibTeX files in a text editor and shell, even if I'm only generating BibTeX files.

The author discusses using non-standard fields like keywords to store extra data. I would recommend that to store additional context about the document. Zotero can store even more than that, including web pages, files, and additional notes. I like that Zotero saves a snapshot of the journal article web page in most instances. Journals sometimes do go offline, so it can be nice to have the web page. I often have detailed notes about particular documents in notes in Zotero. Yes, you can do that with BibTeX, but I could see the extra fields cluttering the BibTeX file. (Contrary to what the author states, the note field probably should not be used as they describe because it's printed in some/most? bibliography styles.)

The search in Zotero is more powerful than grep. Try returning bibliographic entries where one field contains X and another field contains Y. I think you can probably come up with a grep solution for that, but it's so convoluted that it's probably never used in practice. You could use a BibTeX searching program like biblook to get around this problem.

Don't get me wrong. Zotero is far from perfect. It can be slow, and at this point I would prefer a TUI reference manager. But overall, it's the best option I've tried.

> First I used Claude Code to generate all the BibTex files for the PDFs

I think this has an unnecessary risk of hallucinated bibliographic data. For anyone doing something similar in the future, it would be more reliable to make a LLM generate a list of DOIs and have Zotero import the DOIs.

I would suggest looking at more powerful file managers. I use Midnight Commander, which isn't perfect, but is far more powerful than what is provided by default and is also extensible through things like its user menu. My organization system makes heavy use of various Bash/Python scripts and Midnight Commander user menu items to make the file system work for me.

***

A friend of mine has a similar single folder PDF organization system and lately has been trying to better organize it. As I understand it, he's keeping the one folder, but using a reference manager software to track metadata like tags. In contrast, for about 15 years now, I've stored various documents (mostly PDF files) in a folder hierarchy with Zotero for bibliographic data. The PDF filenames match the Zotero citation keys so I can easily switch between the two systems. I'm currently at about 39K PDF files in 4.5K folders with 17K symlinks.

My friend has slowly chugged along, organizing perhaps hundreds of files manually per month. (Maybe he's stopped, I don't know.) It seems to me that transitioning from a one folder approach to something with more context (whether metadata or a folder hierarchy or whatever) is really hard. The best time to add that context is when the file is obtained as I might not remember later. I personally don't trust a LLM to spot the nuances that matter to me when organizing things. A generic organization scheme is of no value to me. I don't know what the goals of your organization effort would be, but this is something to consider.

I think part of the problem is that the people hiring are looking for shortcuts and won't accept any approach that requires more than a certain amount of time.

I've read comments about how X company got hundreds or even thousands of applicants for a position, but they can't possibly look through all of them. Well...

When I worked at the USPTO as a patent examiner, I never had that few documents to search. If I said that I had only 1,000 documents to search, my supervisor would probably be suspicious that my search was poor quality. I was able to find prior art within the first couple hundred documents searched at times, but that doesn't mean that the search could stop at the USPTO. Stopping early like that can come back to haunt an examiner later on when the attorney revises the application. There were many applications where I closely examined hundreds of documents, often looking at drawings and/or text for particular features with a list checked for every single document. I can think of one water heater patent application where I probably closely examined thousands of documents looking for a particular shape of a component, and eventually found it in a fairly obscure Korean patent!

You can complain about the USPTO's quality all you want, but they're at least doing a better job than HR, who simply gives up and says they can't find it when it's right there if they actually put in the effort. The USPTO rejects nearly everything with prior art in the first action.

I'm a mechanical engineer who has written similar tools for work and hobbies. Producing pretty pictures does not mean that the model is physically accurate. Unfortunately, such tools seem be evaluated much more on flashiness and not on more reliable and objective criteria like physical accuracy based on verification and validation test suites. I'm seeing that in the comments here. I don't think LLMs make what I do irrelevant, but I have thought that I'm going to have to improve how flashy my simulations look to compete better with non-experts who use LLMs.

Unfortunately, running an online forum has been a pain for a long time. You have to promote the forum to keep it active, deal with various bad actors (historically spammers, trolls, and black hats), maintain the forum, deal with drama, etc. It adds up and largely isn't appreciated. People act like a forum exists by itself, but it doesn't.

And in the last 5 years, running an online forum has become more of a pain given how badly behaved some bots are. Just recently, I installed Anubis on some online forums I run, and I've been amazed by how much traffic dropped. Before, server load was becoming a problem to the point where all the forums I ran were taken offline by their hosts! I have been thinking about how there's a need for a forum software which produces static HTML for the content, while all dynamic components are behind a login. Bots won't increase server load much in this case. If the forum administrator decides to end the forum, they can easily keep hosting the content without any future maintenance beyond paying the bills. Two of the forums I run are just archives at this point, and I'd love to be able to flip a switch and make them static HTML... (I probably will adapt some script I found on GitHub do to this in the future, maybe with help from LLMs.)