Don't laugh. I am really, really excited about Bash and vanilla PHP.

I landed myself in a shop that sells, essentially, tiny locally-networked systems of specialized Linux boxes. Yes, our core offering is running on the JVM, but all along the edges there is just so much Bash and PHP.

I don't think I'll ever work in a place again where I have this much opportunity to become a genuine old school shell and webshell wizard again. I want to master the Primordial Arts, their endless exceptions to exceptions, and come out the other side as a true master of something completely fucking ridiculous.

I did it once with Microsoft PowerShell. I can do it again.

I once pair programmed (w/@tobinfekkes) a file browser in PHP and bash. PHP would receive the request, exec some bash scripts to cd and ls it's way around the server, then it would output those file names as html buttons to a web page. It was a fun. Bash and PHP are the frowned upon bastard step children of programing languages, but somehow I wish I had your job

Sounds like a dream come true! Any jobs going?

Maybe, if you're able and willing to move to Finland!

Pay is probably under 50k a year though. Can't win em all.

Thank you, but I'm afraid I can't move.

I'd love to hear more about the product though! Without giving too much away, what does the product/company do? :)

Atomically precise manufacturing, that is tech for building structures atom by atom. This is starting to be done regularly in labs. Machines for making single atom transistors are a commercial product now[0]. Forming covalent bonds at desired locations through positional control of reactants has been demonstrated[1]. This is potentially more scalable than the aforementioned approach. ML seems to be enabling too[2].

[0]https://www.zyvexlabs.com/apm/products/zyvex-litho-1/

[1]https://www.nature.com/articles/s41557-021-00773-4

[2]https://www.nature.com/articles/s44160-024-00488-7

We're still barely scratching the surface of AI/LLM/ML. It's correctly getting plenty of attention because it is general purpose technology (steam engine, electricity, the Internet, etc.).

Lot of interesting things happening with NeRF, Gaussian splatting, and similar. We're close to be able to do full 3D video capture of locations and high-fidelity playback/exploration in VR.

Do you think LLM/ML is on the level of steam engine and electricity: a century and more in value, and still holding on; or more of cloud computing level: OpenAI being analogous to AWS, providing loads of value but still on the 1 decade scale?

Much much bigger than cloud computing. Cloud computing is more like steam engine to the internal combustion engine. ICE allowed for airplanes, but not as big a deal as moving from animal power to chemical power. AI is closer in significance to the transistor. Before the transistor, you could do the same things with vacuum tubes. The scale of the transistor made so much more possible. You can accomplish a lot of what AI does with millions of hand-written rules, but AI automates the generation of these rules.

Love that comparison!

I'm so glad that some companies/people are building and releasing models to the public. This amazing development being exclusive to OpenAI would've been a terrible situation for the world.

Are you doing anything with LLM?

In my opinion, for the vast majority of people, and hence for society, we are past the point of diminishing returns when it comes to truly improving the world with technology. We should now switch from unfettered development to trying to be more sustainable with the tools that we have.

Academia has become a game and tech development has become driven by pure consumerism in the race to replace human relationships with the experience of "The Product" so that we have to continually upgrade to maintain any semblance of social cohesion.

There is very little left in tech development that is good. We should think about being more sustainable instead.

The Amish decided that a few hundred years ago.

Tech not invented but popular in sci-fi like transporters or replicators could end hunger or change how we move. 3d printing has reach a general price point on par with other manufacturing but when it does it could change things.

Plenty left that is good and necessary. Don't give up because llms and social networks are overhyped.

Agree with your point about tech, and I don't know a lot about academia to disagree.

One of my issues as someone who works in tech is that it so much capital has poured in and so much attention was on the field that there are relatively few innovations where one can create value. OpenAI being an exception that also proves the rule. A true innovation that literally needs billions of $ to have a chance of happening. Microsoft or Amazon didn't need that kind of money to build successful companies on the back of new business models.

This is driven by incentives though...

The same incentives that enable google to exist, then push the same company to be "less good" by prioritizing growth and public market interests.

As long as you borrow capital, there is always a cost on incentives.

Worse than diminishing returns, today we should expect a new invention to be used against consumers rather than improve the world for us.

Consider the DVR. It transformed the experience of watching television before cable companies figured out they needed to acquire it and worsen it. The damage (to them) is done, we got used to fast forwarding through commercials and now that's a necessary feature with cable.

That could never happen today. Profit, advertising, and surveillance are too strongly built into the process of invention right now. Technology companies have mastered enshittification on a level that older industries never innovated on their own.

People say that I'm cynical but wow this is just off the charts. It also makes me really optimistic that something big is right around the corner. When people start with the "it's all been done" and "nothing good or better will ever happen again" it just means they can't see how things will change. Because they're blind they won't see the next big thing creeping up behind them and when it finally does become apparent they'll exclaim, "Wow what a breakthrough" when the truth is more likely that it's been something that has been brewing for decades. I have some guesses about what the next ones are.

No one said it's been done. There has been healthy debate about the fact that there hasn't been any (major) new social media networks since Snap (and Instagram before it). The cloud computing paradigm is now mostly dominated by 2 players (with GCP a distant third). here are 2 major desktop OSes, 2 major mobile OSes. etc.

This isn't proof that it won't happen again, it just says that the rate at which new innovations in software (and a new workout app doesn't count) has slowed down. LLMs being the exception but it's not like you and I can start an AI company in our garage.

This is a joke right? No new major social media networks since Snap? Eh what about TikTok?

Got me with that one.

The cloud paradigm is for players printing money that need a scalable way to catch all of those dollars by paying 1000 times more for compute on demand.

Create decentralized software and get ahead for the next shift.

When I look a people older than myself, I see a pretty clear stratification of technologies that each age group "just doesn't get"[1], yet it is hard to predict what this technology will be for myself.

My best bet is medical tech and especially body modification and enhancement. While I have been excited and unafraid of all new inventions my whole life, this is a topic that makes me more uncomfortable than excited - a reaction I often observed in people older then me with other technologies. So I'd say invasive body enhancements and augmentation is a good candidate to be the next big thing.

[1] There are always exceptions, but the overall picture is pretty obvious.

Interesting. I am as far from the medical field as could be. Will need to investigate that for sure.

[Almost] anything that mostly/only exists IRL

Software is an enabler - when it becomes the focus, I think something has gone massively awry

Space industry, particularly mining, raw material processing and construction.

I assume you're talking about space junk cluttering up those orbits.

Upscale those laser zappers which are now being fielded to combat drones and put them in high-flying aircraft or on mountaintops. Start zapping the debris and clean out those orbits. Larger pieces of junk can be collected by some iteration of MegaMaid™ (in essence an orbiting Roomba). Problem maybe not directly solved but at least postponed to be either solved this way in a number of years or naturally through orbital decay of circulating junk.

Go to a Lagrange point, the moon or the asteroid belt before that happens.

This is the first time I’m reading of this, why?

I was linked to a paper on this just yesterday.

> The Economics of Orbit Use: Open Access, External Costs, and Runaway Debris Growth (Rao, Rondina, upcoming 2024).

https://www.journals.uchicago.edu/doi/abs/10.1086/730695

And for an arvix/working paper link: https://arxiv.org/pdf/2202.07442v2.pdf

> In our main calibration, Kessler Syndrome can emerge anytime between the year 2040 and the year 2184, with the precise date being very sensitive to the calibration of autocatalytic debris growth parameters.

What does this mean?

A bit ago (7 months) I asked a related question, that may have ideas for fields or technologies to be excited about working on.

https://news.ycombinator.com/item?id=37473795 (Ask HN: Tech that seems to have vanished?, 189 comments)

Memristors, alcohol batteries, laptop frame heatpipes, flywheel energy, smart textiles, and several others were mentioned.

Interesting. Will definitely check out that thread.

I'm a boomer, so I see things differently than younger generations. Nevertheless, I think that solid, heartfelt advice to young people looking to advance in this rapiy changing environment should consider the following:

1. Become an expert in using and augmenting AI tools to accomplish your work, whatever that may be. There is a difference between average and expert in the results that can be obtained. AI expertise will be important at least for the next 10 years.

2. Commit to lifelong learning and adaptation. The constant for the remainder of this century will be continuous, exponential change. When change has a slope near vertical, this should be apparent.

3. The closer you can be to hard science and technology in education and work, I believe the better chance you will have to make significant professional contributions. The CEO of Nvidia said as much in a recent interview.

4. Make a plan by looking at every professional career option and ask this question: "what is the likely result for this field when AGI becomes real and commonplace?" Then choose to examine more closely those fields where you believe you can thrive in the midst of change.

I hope this is helpful, and enjoy the journey!

Regarding your point 3, the challenge is that at present, so many incentives are aligned against that.

I'd love to do hard science, rather than my current meaningless consumer webapp job. But it'd mean... giving up 5-6 years of tech salary to go back and get a phD, and then post-phD my job prospects would be geographically limited to a couple locations where all the science/research jobs are, with worse hours, worse flexibility, less stability and drastically less pay.

It's hard to say "I should quit my fully-remote, wfh, $200k, low-stress, unlimited PTO software job to work in a lab in Boston for 50 hours a week for $90k". I don't have any solution here, kinda just venting through my golden handcuffs.

> my golden handcuffs

I never got this frame of mind. You chose to wear them. You like your golden bracelets. You can take them off any time you want. You are making the choice of staying. At the very least assume the responsibility.

I'm not sure where your confusion is, maybe it's just the terminology. I'll address your points individually and see if maybe that helps you understand what people mean by the "golden handcuffs" framing.

>You chose to wear them. Correct. Because it is the least-bad option that I see as a possibility.

>You like your golden bracelets I do not. I dislike my handcuffs. I just recognize that the other handcuffs I'm allowed to replace them with are worse.

>You can take them off any time you want. You are making the choice of staying I can replace them with different, worse handcuffs. I'm unhappy about being in handcuffs, but if someone suggests I replace my golden ones with the sharp, rusty ones, I will decline and choose to stay in my current ones.

>At the very least assume the responsibility I am? In fact the degree of agency here is central to the whole point. If you tell a prisoner they can choose between death by lethal injection or death by being burned at the stake, they can make the choice of lethal injection and "assume the responsibility" for that, but it's not like it's what they actually want-want. It's what they want under the constraints of the circumstances.

Loss aversion is way stronger than potential but not guaranteed gain.

Unfortunately my college days are behind me, but I'm still learning (luckily most of my jobs have involved me learning about a field and about technologies that I didn't know before).

I wish I had paid more attention in maths class when I was college tho, but it's never too late.

AI sure seems like a good bet. I am also keen on what other fields are out there. I've been trying to understand what are the areas where there's something brewing and will likely breakout in the next 5-10 years.

I don't like playing prophet, so this is a very personal hunch: I believe we're at a social dead end. Whichever new trends we'll see in the next twenty years will come after some kind of mindset switch flipping.

If that sounds too abstract: less safetyism. We could see new kind of vehicles, sports and space exploration.

>We could see new kind of vehicles, sports and space exploration.

So long as we ditch "safetyism" - OSHA, et al, have their place ... but they have become massive bureaucratic nightmares no one can fully "comply" with, and whose regulations are a confusing mishmash of contradictory 'advice'

I really enjoy automation and ETL/ELT pipelines. I got an early taste of AI while studying data science and working with NLP. But I find the automation of tasks, writing tests, and creating tools to help people do their jobs faster, very fulfilling.

PROGCRYPTO (programmable crypto) The capabilities of multiparty computations, fully homomorphic encryption and zero knowledge proofs are super interesting and will allow for new ways of using the internet. In the same way it was hard to imagine what the consequences of asymmetric encryption was at the time when it was conceived, it think the same is true for MPC and FHE today. These technologies are only recently getting feasible. Example use-case: Look for relatives (think ancestry) based on your DNA sequence with cryptographic guarantees, that everyone's DNA stays secret.

Bioengineering artificial organs.

Space travel to Mars and space colonization/mining.

Brain-computer interfaces.

Space, EVTOLs, supersonic transport, robotics

Exciting stuff!

Networking capabilities for WASM could make it the ultimate polyglot platform.

As long as they don't succumb to the pressure to make it Posix compatible by throwing away capabilities, you're right.

Nothing wrong with "Posix compatible"

If you are "Posix compatible" plus 'This Other Thing' ... you have the makings of a great platform

I think the obvious answer right now is anything related to AI. But I'm actually quite excited about the embrace of open-source applications that traditionally have not been, such as Cal.com and Mattermost.

I agree that open source, or even more specifically open and interoperable protocols between applications/sandboxes is where the next leg of software growth will happen on the internet. I don't believe increasing centralization is a trend but rather a swing of a pendulum that would start inverting soon. It will be less about open equivalents of centralized/closed products and services but just the birth of something new with openness and interoperability at the core.

Open source is trending up again these days. Lots of companies (like the ones you pointed out) are trying to be the cloud hosted offering of something that's open source. Love to see it!

I think that technology that will help people to save energy and other resources has a potential to become very useful and valuable. Maybe for single-family homes, but I think the savings could be even greater for multi-family units and commercial real estate. If we can use less energy then we don’t need to build as much solar/nuclear, or, instead use that energy for something else.

Technologies related to automated theorem proving.

Mind sharing why you consider this exciting, please?

Current trendy AI technologies are great for producing solutions but not good at confirming they're good solutions. So, there's increasing need for ways to confirm the solutions. Theorem proving is a way to do this.

Some of the end goals could be: (1) program verification becomes something that's done by default, (2) the entire mathematics literature becomes formalized and verified (and all errors there found and corrected), with programs able to read and formalize papers from the old literature, and (3) new mathematics is done largely by computer, with people acting as "is this interesting?" filters more than agents of discovery. I don't see any reason why mathematics should be immune to domination by computers, any more than (say) Go (the game) was.

> the software industry, and in particular SaaS, has seen immense growth and it's unlikely to see that kind of growth again

Inshallah. I hope you're right.

> interesting fields and industries that might see huge growth

The porn industry

>The porn industry

Honestly ... I think the rise of [almost breaking the uncanny chasm] AI-generated images and video is going to both help and hurt the porn industry

How will it 'help'? Generate, on-demand, whatever form of sexual fantasy the user can prompt

How will it 'hurt'? Generate, on-demand, whatever form of sexual fantasy the user can prompt

When actual, IRL humans are involved, there are always limits to what can be produced

When that limitation is gone ... the cuffs will be off!

Is AI-generated 'underage' sex still "pedophilia"? When those 'engaging' in whatever activity the prompter has suggested 'do not `exist`'?

What about so-called 'extreme' BDSM/pain/fetish/etc? If the 'actors' are all AI-generated ... is it still [potentially] `wrong`?

There are a HOST of ethical and legal issues that are about to hit the porn industry ... and those who want to - rightfully - prosecute those who want grossly-illegal/-unethical material!

Strictly by the letter of the law ... it 'cannot' be `kiddie porn` if the images are not of real people.

Yet ... by the spirit of the law ... it most certainly is `kiddie porn` if the images are of obviously `underage` individuals!

Or what about so-called 'snuff videos'? It is actually a beheading if the `beheaded` is AI-generated?

The legal system is going to have a LOT of work in front of itself in this dawning era, to be sure!

Web apps, that thing that Apple killed recently.

Yes but then they changed their mind! https://9to5mac.com/2024/03/01/apple-home-screen-web-apps-io...

I'm not really excited about any technologies since I'm sure anything new will be expensive, complex, break easily, be of marginal use, or come with privacy issues, etc.

I wouldn't look for a specific industry or anything interesting. It's all pretty much corporate BS. Take whatever job pays pretty well and isn't obscure tech so you'll be able to switch jobs. Finance, law, and institutionl sales (gov, schools, hospitals) aren't going anywhere anytime soon.

Neuro-Symbolic AI

Portable, a tool that runs anything anywhere

so...a 'perfect' emulator?

Whatever the kids are really passionate about.

In other words, girls and boys. May they find the joys to be had when the two sexes meet - or for those few who have different tastes, the same sexes - and may the former consider the natural result of that union to be a boon for them and for society. Go forth and fork(2), ehhhh procreate.

To be clear here, your answer to "What fields or technologies are you excited about?"

Is "kids hooking up with each other?"

Yes, in a field. Try it, it can be quite exciting and certainly helps to create a better future. I'd even go so far as to state that without that there is no future.

Can you expand on that? Are you an adult and where was the field you had this 'quite exciting' experience?

This not being an adult outlet I will not go into detail about which fields are suitable for such activities. What I can say is that you nick - CyberDildonics - is apt for one who asks this question and that the activities alluded to tend to negate the need for anything related to it, especially when it comes to the Cyber part.

So, Roblox.

Or Minecraft :D

Legos!

Tiktok

AI is a personal tutor, amd the biggest change in education since the invention of writing.

I imagine a day when it is socially unacceptable to hire someone to do something for you. You have AI and Amazon, why not learn, order the parts, and do it yourself? Imagine a world where we all work a few hours a week but occupy ourselves learning new skills and applying those skills to improve our environments. This would also keep us ready to be useful at work.

I am reminded of a famous quote: "A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects." -Robert A. Heinlein

I am afraid that AI tutors and teachers will slow down our progress in ethics and moral issues.

Model X trained on all our data thinks Y is ok and that permeates lessons here and there. Raising a generation that does not rightfully challenge Y as it should.

As a somewhat stupid example: had LLMs been invented in thr 1800's would they teach whatever topic with an underlying sentiment that slavery is a just and natural state?

Teachers and tutors have always been an influence in the formative years of younger minds. I am afraid of substituting that human influence bringing intellectual and moral challenges for a one-size-fits-all frozen in time pasteurized worldview.

To be clear: LLMs as educational tools; inevitable. LLMs as teachers - no, please

Let me know if this is a stupid feeling (and a feeling is all I have) or if it is too much of a Old Man Yells at Cloud thing

You make a good point. Something I hadn't been able to put my finger on but which was bothering me on a subliminar level.

It is likely that a main LLM will dominate the play field (network effects) and it will breed a thinking monoculture. The more entrenched it becomes, the less resilient society will be, because we will become over reliant on its model weights.