maxander

It’s not that dumb- if a human gets exposed to space the water in their exposed tissues will boil off, leading to evaporative cooling. In a vacuum, evaporative cooling can get you ~arbitrarily cold, as long as you’re giving up enough fluids. I don’t know whether you freeze over or dry out first, but I’m sure someone at NASA has done the math.

This is the basic concept of my favorite computer game back in the 90s (!): https://en.m.wikipedia.org/wiki/Creatures_(video_game_series...

The interesting thing about this blog post isn’t the point it makes (which was endlessly made about cafes in all blue-state U.S. cities in the 00s) but that it’s an Iranian writing in English, and thus for an Western audience, about Iranian culture.

What is going on with that? I don’t even dare to speculate. But something more interesting than bougie coffeeshops, certainly.

Sadly, I think the modern alternative for ncurses-based interfaces is javascript.

Yeah I was unclear there - I meant the other use of the & operator, 'foo & bar' where bar starts after foo completes. AFAIK, if you ran foo by itself, there's no way to make bar automatically run after foo completes.

My primary project the past couple weeks has been applying the AI interpretability technique from Anthropic's famous paper this past summer (you know, the Golden Gate Claude one) to single-cell RNA-seq data. What works on one huge, inscrutible vector space ought to work on another, right? In either case, a fun way to keep in practice, both for comp bio and deep learning.

My side-project, in essence, the bash '&' operator for cases where the first process is already running. It took me months of searching before I could believe that this doesn't already exist, but there you go. I gave in to feature creep, of course, so it's a bit more than that now (I made a ncurses-based dashboard? Why??) but someday soon I'll make it public.

Location: Boston, MA

Remote: Yes

Willing to Relocate: Unlikely

Technologies: Python, R, Javascript, AWS, GCS, Azure, SQL, pandas, numpy, scipy, Tensorflow, Keras, [Py]Torch, SQL, Nextflow, many others

Resume/CV: http://alexander.bio/resume_2024.pdf

Email: max@alexander.bio

Data scientist and engineer with over a decade of experience centering around biological data analysis and pipeline development. I am equally at home discussing pathway enrichment analysis results from an RNA-seq dataset, as debugging the loss function in a deep learning model, or as architecting a multi-stage processing pipeline on any of the major cloud service providers.

SEEKING WORK | Boston, MA | Remote or on-site

I am Max Alexander, experienced data scientist and bioinformatics engineer, available for short- or long-term contracts. My background is primarily in data analysis and pipeline development for biological data (genomics, transcriptomics, proteomics) and I also hold a Master's in artificial intelligence.

--

Email: max@alexander.bio

Site: https://alexander.bio

LinkedIn: https://www.linkedin.com/in/w-max-alexander/

Github: https://github.com/maxalex

I don’t know whether this is a provocative or tedious thing to say, but the quintessential ‘enlightenmentware’ to have come out of the past several years is ChatGPT. Name anything that brings as much functionality with so simple an interface and so elegant a core!

(It’s simply a pity that we can’t install it locally or tinker with the internals. :) )

The avenue towards making this a viable, routine cure for HIV is by performing a “transplant” of the patient’s own immune stem cells that have been genetically modified to carry the anti-HIV genes. That avoids (if done right) the horrific autoimmune challenges you describe, but like everything else it comes with its own technical challenges that are still being worked out. (I think https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4935568/ is a good review of the whole idea.) To me (though I’m not an HIV researcher by any means!) this sounds like the most promising shot we have at regularly curing the disease.

I’d recommend mynoise.net ; most of the sounds are for-pay but white noise and a few others are free, and it includes an equalizer (and some presets for pink noise and etc.)

An app, though, not a website.

It’s not news to anyone that NATO could win a (conventional) war with Russia. The important result is that any sizable chunk of military force in Russia itself stands a good chance of it.

SEEKING WORK | Boston, MA or remote

Open to freelance/consulting work in bioinformatics. 8 years of experience in various areas of bioinformatics and computational biology, with particular expertise in the analysis of LC-MS proteomics and transcriptomics data. Well-versed in taking a problem in the language of biology and finding the appropriate computational solution.

Linkedin: https://www.linkedin.com/in/w-max-alexander/

Email: max@alexander.bio

Is this true? Flat rectangles pack more efficiently than tubes. With mattresses, afaik, the big difference was vacuum-packing reducing overall volume, not shape.

It’s definitely not too old for anything, if you want to do it. Source: I transferred from local college to better college at 20 myself.

There will be plenty of 20-year-olds where you transfer, and no one will care what year you technically are. Don’t worry about that. College is not like high school.

As other commenters have noted, the big question is whether you want to spend the extra time in school (and spend the extra money for the extra time - college debt is serious business!) But you know your own situation best, and an extra couple years of classes can be spent profitably if you’re thoughtful about how you use them.

But, here’s another possibility to consider- finish undergrad at your current place, and then go for a Master’s degree at the more prestigious school? It’s an extra two years in either case, and then you have a whole postgraduate degree as well.

I recommend Matt Levine’s blog Money Stuff [0] for analysis of just how bizarre Musk’s twitter saga is relative to usual day-to-day corporate acquisitions. I also recommend his twitter thread just now breaking down his reaction to today’s news [1]

[0] https://www.bloomberg.com/opinion/authors/ARbTQlRLRjE/matthe...

[1] https://twitter.com/matt_levine/status/1545519476707319808?s...

Respectfully, the perfect is the enemy of the good, and it’s entirely reasonable to publish what you have now. If later you make further improvements, you can simply publish again.

A bit of black electrical tape, carefully applied, is pretty discreet on a typical black casing

There’s no sign that you’ll be able to turn it off in the car ahead of you, though.

Energy consumption would be limited by supply and energy dissipation, and energy dissipation (roughly speaking) depends on surface area and thermal conductivity. So, yes, for all devices made with a given shape and given materials, there is such an equation.

In theory, I guess, there’s also such an equation for the “upper limit” of heat dissipation for size, for a computer using the optimal shape and optimal materials. But since that probably concerns several-atom-thick devices made of diamond or something, that’s probably not practically useful anytime soon. :)

Indeed, Dyson spheres would need to release some amount of waste heat, due to the laws of thermodynamics. If there were too many of them that heat would be visible (at least to infrared telescopes,) so we know that’s not the explanation for dark matter. Same with dust and other non-luminescent matter- it’s all way more visible than dark matter has to be, somehow.

It’s a nice idea; perhaps something along the lines of a mix between Kerbal Space Program, SimCity, and Satisfactory. There’s a big problem for making this a practical step towards Martian colony planning, though; making a simulation game accurately reflect the real world is hard! If you’ve played any of the three games I mentioned, and have even general knowledge of the fields they simulate (rocketry and space travel, civic engineering and economic management, and manufacturing, respectively) you’ll see immediately that they misrepresent reality to a huge degree, and that as a result good ideas in the game don’t at all translate to reality. (Arguably, KSP reflects its subject matter best, but it’s still limited enough that discussion of the game often involves talking about “very Kerbal” designs that would be outright absurd as real aerospace engineering.)

Even if you can map out the elements of a Martian colony’s technological, social and economic underpinnings, having a game reflect those elements realistically is likely to take a heavy toll on its quality as a good game- simply because the world is very complex. There’s only a limited amount of complexity even the nerdiest gamer wants to deal with. If you doubt me, try playing some of the older Paragon Studios games- and even then, e.g. Victoria 2 was a downright impressionistic take on 19th century geopolitics. An accurate rendition of a 21st Martian colony would be far worse. Your player base would be limited to extremely wonky enthusiasts, and those people are likely already writing papers on the subject regardless.

Regardless of all the above I like the idea and have thought song similar lines myself in the past. But I can see why it doesn’t stand out to Musk as an important use of his time. He seems to feel someone else will come along to solve the actual colonization problem, as long as he just provides the rockets.

If you want to not be behind a screen, what you’re looking for might not be mechanical engineering at all. The guys/girls on the floor of a shop full of CNC machines and 3d printers aren’t usually engineers as such. Manufacturing/fabrication is obviously a whole different universe than the typical HNer inhabits, and has a very different culture- but it can be very high-skill while usually a having lower formal training requirement. (Perhaps even can be a useful stopover before beginning studies to go back behind the computer screen as a MechE.)

I don’t know, but there’s always a relevant XKCD, even for XKCD: https://xkcd.com/1159/ :)

I've been interested in this field long enough that I'm still plenty happy with the fact we now have a longevity biotech field, despite it's issues. :)

I agree with the author that longevity biotech is frequently tripping over itself with enthusiasm to bring interventions to market (or some variety of off-label use) well before they're on solid scientific ground. You don't see this in, e.g., cancer research- if a novel kind of anti-cancer drug tests well in mice, you (almost) never see a sudden cottage industry pop up around it the next week. Obviously part of the difference is that most of us aren't expecting to suffer from cancer in the immediate future, whereas everyone expects to age; aging research faces a lot of weird psychological pressures that most industries do not.

For all that, though, there are a lot of parts where I think this article is too skeptical. As someone who uses machine learning in (non-aging-related) biotech, the discussion of "garbage-in/garbage-out" put me in mind of "Yes We Have Noticed the Skulls" [0]; we're well aware that data quality coming out of molecular biology is a problem, and anyone working in the field could give you a long list of the kinds of biases, distortions, and gaps that are known to occur in their areas of interest [1]. Not all of the grand claims made about biotech AI/ML are taking this expertise into account, certainly, but certainly many of them are.

[0] https://slatestarcodex.com/2017/04/07/yes-we-have-noticed-th...

[1] The list, regardless of what variety of data it applies to, would always end with "sometimes something unknown happened and the data just turned out bad, oh well."

Effective science communication has many forms! :)

Most likely, a user of the New dialog was shown “right-click here and select this and now ‘Word Document’ and then…” and thenceforth has followed these instructions essentially as a mystic ritual which achieves a necessary result, without a real understanding of what the steps mean. They don’t know that they’re running a program, they just know they’re making a document appear on the screen. Though they could probably guess the phrase “Microsoft Word” is pertinent to the ritual, they don’t know what it refers to, so transferring that insight would be a matter of guesswork.

You could just as easily show them the steps on the Start menu, it’s true- but then you have the separate question of saving the file somewhere and being able to find it again later. I know from helping elderly relatives with their computer problems that “directories” are another confusing concept for a lot of people. Pointing to the Desktop and saying “save it here” makes intuitive sense and makes the file easy to find later, but to e.g. “Save As” to the Desktop from inside Word requires an understanding of the occult dual nature of the Desktop-as-a-directory and the Desktop-on-the-screen, which is an insight reserved for enlightened wizards. Using the New menu lets the user restrict all their interactions with the file system to, at worst, Windows Explorer, which is a huge simplification!

(Example of where this is coming from; once when my father, a practicing medical doctor in his 70s, needed to e-sign a PDF on an unfamiliar computer I spent about half an hour trying to explain the concept of “close the document in this app and open it in this other (e-signing enabled) app.” I could about as well have been speaking Greek, or telling a Flatlander about flying.)

> I can't see how the New menu is a particularly good solution because it doesn't seem any more discoverable than the list of document-authoring programs in the Start menu.

I think you’ve missed the key point of this article, which is that a lot of people use computers without a clear idea what programs are. They know they can double-click a spreadsheet file and have a window open that lets them do things with that spreadsheet; but to them, I guess, that’s just “what a spreadsheet file is like” and there is no notion of this thing called Excel which is mediating their experience of the file. To a person like this, “discovering the program in the Start menu” is not something that can happen within the bounds of how their ontology of computers works.

It’s a credit to the designers of Windows that they design to accommodate users like this, though it also goes a long way to explaining why Windows is so annoying to many professional computer users.

> So I guess they never got around to implementing adoption.

If you think about this, it seems pretty obvious the reason they didn’t include adoption was that it’s overpowered. The CK series is all about managing dynasties, so the randomness of reproduction is a major game mechanic, one of the primary sources of problems and challenges. Adoption would totally negate all of that.

Like, if Henry VIII could just have adopted a son, his whole story would have been so much more straightforward! :)

(None of which really justifies the Paradox response, of course, since adoption isn’t necessary for having same-sex marriages (even player ones) and not losing.)

If we have full-on adult genetic modification capable of the, ah, dramatic example you provide, we’ve certainly figured out a way to get around in-vivo test difficulties. For better or worse, any biomedical advance comes up against that problem sooner or later.

Therapies to slow aging have the particular problem that it could intrinsically take decades to show an effect, sure- but that’s simply reason to be a bit more ambitious and aim for therapies to reverse aging, which could be tested rapidly in already-old patients. :)