Well, maybe you should just try for the hell of it and see how far you get? Becoming fit seems impossible to a morbidly obese 45 y.o, and it is if that person's expectation is unreasonable, but if they just change it to be more reasonable, break it down into manageable routines, then they can get somewhere eventually.
Find some papers, fill many gaps, dedicate a few years in your spare time, in 6 months you'll be 6 months closer than you were.
Whether there's a reason or not, idk, it's something to do, be curious. Don't forget that by dedicating their life to something, they're naturally not dedicating their life to other things, things that you might be able to do, like climbing mountains, making pizza, or coming up with witty banter in social situations.
Not to be morbid, but…in 6 months you’ll also have 6 months less of your life left to do those things.
In the past few years, I have felt—-whether due to my aging, the chaos of current events, the mindless march of technology, who knows—-that our time here on earth is a gift that we squander at our peril.
I think some would find that bleak and grim, but I consider it an existentialist battle cry. Make a list of those things that someday you’d like to say you’ve done, or know, or have built, or have given to others…and do one of them.
The time passes anyway.
All the more reason to spend it on things that matter to you. The opportunity cost of six months spent deep in abstract CS papers is six months not spent teaching your daughter to play guitar, visiting that place you always dreamt of seeing, finally finishing that book on your nightstand, etc.
When it comes to that category of your own interests, I don't really think one can afford not to spend time on them, lest you hollow yourself out. Whether any one thing is worth the time over another, like grinding papers vs travel, they're not always mutually exclusive; although trying to do both in parallel might be silly, I personally like to shift my attention periodically. I'll go and spend a few months learning, and then go adventure. I don't that much, but I'm happy to meet up with friends and do that too, and it means taking time away from video games or learning, and that's important too.
I guarantee thats more time than most people will spend teaching their kids any musical instrument.
Just spend a week mapping out what you do ans how long it takes you every week and I'm pretty sure you can find double digit hours spent somewhere, maybe even right here on HN
For me, four hours a week is sufficient to stay up-to-date on an active research area but making forward progress requires at least twice that.
> You are awake for at least 16 hours of the day, you telling me you cant find 4 hours a week to read a paper? So 4/112 hours or around 3.5% of your week...
Using awake hours as the denominator is misleading because most people have other non-discretionary time commitments besides sleep. For me I'd estimate ~60h/wk sleep, ~50h/wk work/commute, ~30h/wk non-discretionary upkeep of children/relationships/home/body. Assuming 8+h/wk to make progress out of the remaining ~28h/wk of discretionary time means I can handle about three non-discretionary priorities. (Pre-kids I could handle about five.)
Therefore, when someone with a job says "I don't have time" to pick up a hobby, skill, language, outside research area, instrument, volunteer position, etc I don't interpret their statement as meaning it is physically impossible for them to rearrange their schedule to accommodate it. I (and I suspect most people) interpret the statement as them admitting that it's not one of their ~3-5 non-discretionary priorities.
There are legitimately busy people and then there are people who wish they could achieve X if only they had time but don't put any effort into making time for that.
HINT: if research is directly related to your job, allocate time to it during working hours, those aren't 40-45 hours of time a company gets to take from you and also get benefits from your out of work time. I'm reasonably sure your boss would happily let you allocate an hour every now and then to improving yourself as an employee and if they don't, well... The internet has their usual answer to that even though I don't always agree.
Sometimes this is the result of black-hat products hacking their dopamine cycle, in which case screentime or a friend can help. However, I've found that in some cases staying on top of the zeitgeist like this is actually in someone's 3-5 priorities. In that case saying they have "no time" for X is another way of saying that using TikTok is a higher priority than X for them. (Baffling to me, but a valid choice.) I similarly know people who spend a non-trivial amount of time on other "useless" activities like watching TV shows, playing video games, reading novels, learning esoteric languages, growing plants with no utility, commenting on online forums, etc. Who am I to judge if they find it valuable?
So as technically imprecise as "I don't have time" is, I understand why people use the expression. When someone suggests that I should volunteer for a cause, participate in an activity, go to an event, learn a skill, watch a TV show, read a particular book, learn a language, etc and I tell them that it isn't a high enough priority to displace any of my existing priorities, they sometimes get defensive and/or attempt to litigate my current priorities.
> I'm reasonably sure your boss would happily let you allocate an hour every now and then to improving yourself as an employee
Absolutely, this is a major perk that knowledge workers should take advantage of. I'm spending quite a bit more than "an hour every now and then" to learn about LLMs and accessibility because they are in the intersection of my interests and my job responsibilities. However quantum computing (or game design, solar vehicles, gardening, etc) are not in that intersection and would count against one of my discretionary priorities.
I will always try to convince people against mindless media like ticktok, well unless it's in their life goals to be an influencer but that may also be an issue...
Other cauaes though, sure I don't mind if you don't have time to volunteer etc.
At 70 nobody will be proudly say "oh yes I've spent years reading up on this topic!".
Just like I don't know how to build a solar panel or how to do organic chemistry.
However, that's also exactly why I didn't say anything like "You should learn X", because it's just a curiosity, and there's many curiosities. For example, last year I failed an interview at Apple because they got the impression my hardware-level knowledge of computers wasn't there, and it wasn't, and that convinced me to finally try and work my way through NAND2Tetris, which I'm now about 3/4's of the way through, and feel was incredibly rewarding even though the net benefit is likely nebulous. I was out of work then for about a year and a half, and it helped me pass the time well too, in a much more spirit lifting way than grinding through yet another rest api project or frontend framework.
Eventually a project may come along that I'll feel is compelling enough to dedicate some serious time to AI/Crypto, and I'll consider it then, but if I were to just try and learn it for no reason at all—including innate curiosity—I don't think it'd stick.
Not an easy goal, good luck!
I'd use the example of hiking literally all day without the promise of a good viewpoint; I'd invite the person out, with only a plausible estimate of the time required, and they'd want to just find something that takes less time so they can schedule something afterward. Along the way, they'll be rushing to meet that time, because this is just exercise or whatever to them, and in some way they aren't at peace the idea that we're both just here in the forest maybe chatting maybe not, there's no tangible justification for the mission.
Another type of person would replace tangible outcomes with the feeling that they always need to be learning, regardless of what it is, because it's intrinsically virtuous, and they also sometimes fail to be at peace with doing something for no reason, or nothing at all.
I've wavered between these over the years, and now I'll learn something because it's a clear weak spot, or I can imagine how it might be interesting, and if I don't have anything else that's more compelling (including doing nothing) I might give it a go. What's different now than a few years ago is how much I respect serious time involvement. Anything I decide is worth trying to learn is something I need to feel capable of dedicating serious energy to, at least in the first year; if I can't or don't want to, then maybe I won't, and I shouldn't fool myself into thinking I should or will, because I have other things going on. If I'm going hiking, that's my day, that's it, that's the whole activity, if anyone wants to join me then that's great, they need to accept the same mentality or they can stay home. If we happen to get back before the bars close, then that's great too. I might find Swahili interesting too, and if it seemed worth trying, I'd dive in on the basis that I'd just get a sense for how a different language works, and that there might be something surprising along the way, which to me is inherently valuable.
Yes but - it's up to each individual to decide on their own definition of 'squandering'. Ultimately everything we do is in service of our own search for meaning in life, and learning for its own sake can absolutely fulfil that role.
Thank you for writing them.
Well... guess it's time I start learning quantum computing then
Simple in theory, juggling plates and knives in practice...
https://www.quantumplayground.net/#/home
You don't need a "real" quantum computer to mess around with quantum computing and learn how it works any more than you need a supercomputer to play around with algorithms and learn how they work.
Shor's algorithm is from 1994.
You can in general start with these search keywords: qiskit caterpillar yosys.
Thing is, I don't know how to get out without on the one side giving up my comfort zone (well paid, doable work), and on the other side gaining responsibility / being looked at as an expert in any field (that's where impostor syndrome and responsibility aversion comes in). I really need a holiday lol.
Spent yesterday afternoon and this morning learning what I could. I'm now superficially familiar with quantum coherence, superposition, and phase relationships.
In other words, you got this. Now I gotta learn linear algebra. brb.
One thing I did forget to mention is that you can play with this stuff in a "familiar to software developers" way in our VS Code playground at <https://vscode.dev/quantum/playground/> . This is a 'code first' approach familiar to software developers leveraging VS Code integration. The playground is pre-populated with a bunch of common quantum algorithms.
You can also install the extension in VS Code directly (<https://marketplace.visualstudio.com/items?itemName=quantum....>), you don't need to run it in the browser, but even in the browser it has a fully working language service, debugger, evaluator, quantum simulator, package management, etc. It's all written in Rust and compiled to either Wasm for the browser and VS Code extension, or native code for the Python package. (I'm thinking about doing a video on how we build it, as I expect it will interesting to this type of crowd. Let me know if so).
Disclaimer: I work in Azure Quantum on the product mentioned. AMA.
I'm unsure I am even asking the right questions. I'd appreciate any direction you can give me!
What do you hope to be doing in 5 years? Architecting quantum solutions? Reselling or consulting on cloud solutions? Building quantum applications?
I think the quantum space will have quite a bit of progress in 5 years, but I think most experts in the space (of which I'm NOT one) think we're still over 5 years out before there's broad adoption on running quantum programs with significant business value. (i.e, it'll still largely be researchers and bleeding edge adopters).
Opinions are my own, etc.
Linear algebra does seem to be a hard wall I've seen many smart software engineers hit. I'd honestly love for someone to study the phenomenon and figure out why this is.
So, I really agree that it's about timing and curriculum. For one, it appears somewhat abstract until you really understand geometrically what's happening.
So, I surmise that most non-mathematicians don't have quite the mathematical maturity to absorb the standard pedagogy when they first approach the subject.
Yes, but most people, like me, never gain even remotely intuitive understanding of it and quickly forget it after passing a few dreadful exams. Just like with most other math.
If you want to go deeper into quantum computing, I can highly recommend Scott Aaronson's own book, "Quantum Computing since Democritus"[0]. Although I have a background in physics and math, I found his style lively and engaging with truly unique and compact recapitulations of things I already knew (two come to mind: his description of Cantor's diagnalization argument, and the assertion that QM is the natural consequence of "negative probabilities" being real. The later is a marvelous insight that I've personally gotten a lot of use out of).
It's also useful to understand the boundary of what quantum computing is. At the end of the day what we'll see are "QaaS" apis that give us the ability to, for example, factor large primes. You won't need to know Shor's algorithm or the implementation details, you'll just get your answer exponentially faster than the classical method. I would not expect desktop quantum computers, languages designed for them, or general user software designed to run on them. (Of course, eventually someone will make Doom run on one, but that's decades in the future.)
https://www.alibris.com/booksearch?mtype=B&keyword=quantum+c...
https://podcast.clearerthinking.org/episode/208/scott-aarons...
First of all, the formalism/practice gap is real: taking API calls and updating a database correctly has a mountain of formalism around it. And it is not easy to get right! Concurrent sequential processes and distributed systems theory and a bunch of topics have a huge formalism. It is also the case that many (most?) working software engineers have internalized much of that formalism: they “play by ear” rather than read sheet music, but what matters is if it sounds good.
Second, whether it’s quantum computing or frontier machine learning or any other formalism-heavy topic? It’s eminently possible to learn this stuff. There’s a certain lingering credentialism around “you need a PhD” or whatever, I call BS: this stuff is learnable.
Keep hacking, keep pushing yourself on topics you’re passionate about, but don’t consign yourself to some inferior caste. You’re just as likely as the next person to be the next self-taught superstar.
Quantum error correction is one of those “wow” moments like euler’s identity, it is worth making the effort to get there.
Any suggestions for where I can learn more about the theory of quantum computation?
I’ve made a list of 4-5 things at which I want to be extremely good at in 10 years compared to where I’m today. Now I just spend time on those. I occasionally wander into something new just for the sake of diversion.
https://metacpan.org/pod/Quantum::Superpositions
(Sadly, the Perl Module has to do classical calculations underneath to get the Quantum computing code/functions to execute, but it let you experiment with silly little QC toys - "Look, I factorised 15 in parallel!!!")
to me the whole endeavor smells like a bait and switch or something. I remember about 10 years ago canada or someone had at least a few hundred qubits if not close to 1000 of them, but these were physical qubits, and don't represent anything, really. Google's 105 finally makes a "fast enough" single qubit or at best half of a pair.
I'm wondering if it's time for me to switch professions and give up compsci / software altogether.
Many pre-trained models and libraries that hide most of the complexity.
However, my sentiment is rather this: I wouldn't pass an assembly programming interview, but I know enough about it so that I know what I don't know. Same with embedded programming, fpgas, machine learning stuff, big data, networking, etc etc.
As for LLMs and quantum computing, I don't even know the basics, have no idea about the broader science behind it. Worst is that I don't feel like it interests me, I don't feel excited about it.
I guess if tomorrow I had to work with them, I could learn some "libraries that hide the complexity", but it leaves me with an empty feeling about these new technologies. Hence the existential question if I'm "too old for this" career path at all.
About 15 years ago I became interested in really advanced cryptography, because it was presented at a Bitcoin conference I went to. If you think AI is hard, that's kindergarten stuff compared to the maths behind zero knowledge proofs. And because nobody cared at that time outside of a handful of academics, there were no tutorials, blog posts or anything else to help. Just a giant mound of academic papers, often undated so it was hard to even figure out if what you were reading had been superseded already. But it seemed important, so I dived in and started reading papers.
At first, maybe only 5% of the words made sense. So I grabbed onto those 5%. I read a paper, put it down for a while, re-read it later and found I understood more. I talked to the researchers, emailed them, asked questions. I read the older papers that initiated the field, and that helped. It was a lot of work.
You know what? In the end, it was a waste of time. The knowledge ended up being useful primarily for explaining why I wasn't using those algorithms in my designs. A lot of the claims sounded useful but ended up not being so for complicated reasons, and anyway, I was mostly interested in what you could do with the tech rather than the tech itself. Turns out there's always a small number of people who are willing to dive in and make the magic happen in a nicely abstracted way for everyone else, for any kind of tech. QC is no different. There's, as far as I can tell, very little reason to learn it. If QC does ever "happen" it'll presumably 95% of the time be in the form of a cloud service where you upload problems that fit a quantum algorithm worked out by someone else, pay, and download the answer. Just like LLMs are - another topic where I was reading papers back in 2017 and that knowledge turned out to not be especially useful in regular life.
Learn the details of stuff if it naturally interests you. Ignore it if it doesn't. Being a specialist in an obscure domain can occasionally be like striking the jackpot, but it's rare and not something to feel bad about if you just don't want to.
Why should you? I agree with your sentiment, super advanced quantum physics is probably out of reach for 99% of the population (I'm estimating here but I think it's reasonable to assume that's the average IQ of the physics PHDs who can actually understand this stuff to a deep level). You can probably make the effort to understand something about what's going on there, but it will be very superficial. Going advanced quantum physics takes a huge amount of effort and an incredible capacity for learning complex things. And even the advanced physics guys don't and can't understand a bunch of very elementary things about reality, so it's not as if the feeling of not understanding stuff ever goes away.
Here’s a sneak peek from Lecture 6: https://youtu.be/6rf-hjyNl4U
You can sign up via: https://quantumformalism.academy/mathematical-foundations-fo...
> Only an elite few get to touch these machines.
But lately many can run quite a lot of AI models at home. Doesn't require too crazy of a setup.
Why not build something software fun at home that doesn't involve a DB? Maybe using some free AI model?
I did experiment lately: automatically "screenshot" a browser and ask an AI to find the URL and ask if the URL and site looked like a phishing attempt or not. Fun stuff (and it works).
I tried installing one of these "photo gallery" in a Docker container (where you can put all your family/travel pics and let anyone on your LAN [or on the net] browse them). I saw some of these have "similarity" searches features. I also saw that SAM / SAM2 (Meta's Segment Anything Model) was plenty quick: some people are using these to analyze video frames in real-time. So I was thinking about sending all my family pictures through SAM2 (or a similar model: I saw some modified SAM2 to make it even faster) and then augmenting the "similarity search" by using the results of SAM2. For example finding all the pictures about "pool", etc.
And why limit myself to pictures? I could do family vids too: "Find all the vids where that item can be seen".
Possibilities at the moment seems endless: times are exciting if you ask me.
Being part of a highly educated elite group like that has some huge benefits, but they have also shackled themselves to an insanely specialized and highly difficult niche. I can't imagine the stress they are under and the potential for despair when dedicating years to something with so much uncertainty.
You can't know everything but knowing what you don't know and when to rely on someone else to know what you don't know and to confidently quote or use what they know that you don't know, is a skill too.
"Critical thinking", and I suspect you do know how to do that and whilst this blog post might be somewhat impenetrable it is still might be useful to you, even as just general knowledge. Use your skills to determine - for you and you alone - whether it is truth, false or somewhere in between.
Besides, a mental work out is good for you!
TL;DR: whatever you’re doing, there’s probably someone who wishes they were doing it instead.
Source: teaching beginners piano for years. Of course sheet music looks like gobbledygook, until you've spent a bit of time learning the basic rules!
For now, people were saying the same thing about mainframes in the 60s
I will also note that sometimes when I read HN links and think one thing then read the comments and people know enough to take issue with what is being said and even call it out.
React 19 is out and it's time to hit the docs eh to solve the same exact problem we've had for the last 30 years in a slightly better (or NOT) way.