HNHacker News
TopNewBestAskShowJobs

dogcomplex

627 karma · joined August 29, 2016

submissionscomments
dogcomplex··on Move over, tractor – The farmer wants a crop-spraying drone
You're very knowledgeable, this is great. Is there any automated tech on the market you've seen to do the spotting for you, so you can dynamically ground drones as air conditions change? Also, are you piloting all these manually (yknow, if they weren't gathering dust), is it a preset flight plan, or are you already experimenting with AI control? Would think even an LLM fed with air traffic sensor data could make the call to ground drones and switch between flight plans dynamically at this point - though I'd want it thoroughly tested and airing on the side of caution always. Also - what kind of height do you require for drone dusting? Could you get away with keeping it low to the ground, dusting only a few plants at a time at the expense of longer flight time and battery charging, but at least hopefully practically low enough to skirt any potential collision and legal issues, and potentially still entirely automated? Would think that kind of specificity also allows for stuff like dusting individual crop areas differently according to need, too.

All just thoughts from a programmer here - very cool you get to experiment with this kind of stuff in a practical setting. Hope it gets more practical soon!

Bummer the law is so slow on this.

dogcomplex··on Scientists Find an 'Alphabet' in Whale Songs
So what you're saying is to some extent everything's made up and the points don't matter, so maybe we can rope animals into this same messy web of loose promises too!
dogcomplex··on GPT-4o
If you want to talk facts, then those critics are similarly on weak grounds and critiquing feelings more than facts. There has been no actual sign of scaling ceasing to work, in medium after medium, and most of their criticisms are issues with how LLM tools are embedded in architectures which are still incredibly early/primitive and still refining how to use transformers effectively. We haven't even begun using error correction techniques from analog engineering disciplines properly to boost the signal of LLMs in practical settings. There is so much work to do with just the existing tools.

"AI is massive hype and shoved into everything" has more grounding as a negative feeling of people being overwhelmed with technology than any basis in fact. The faults and weaknesses are buoyed by people trying to acknowledge your feelings than any real criticism of a technology that is changing faster than the faults and weakness arguments can be made. Study machine learning and come back with an informed criticism.

dogcomplex··on Did GitHub Copilot increase my productivity?
It is honestly shocking how different people can be, even within programming. Been finding LLMs very useful for pair programming and low-effort spin-ups of projects I would just never do without the tools. Frankly, I'm waiting any day for an architecture to arise that takes lossy LLMs and has them error correct enough to produce 99.9999% reliable progress on simple tasks cheaply. This feels incredibly close, and if the unit of intelligence of each step of those is around GPT4 level? Jesus christ. Programming massively automated overnight.

Any programmer that is this dismissive of these tools frankly... hasn't been using them right, and has no imagination - or a massive ego. This stuff is still far too early to make such judgements.

dogcomplex··on Scientists Find an 'Alphabet' in Whale Songs
We'd need contextual tracking of what the whales are actually doing/communicating to match to the songs. An LLM would be excellent at finding any correlated patterns between the language and actions, and then mapping those to similar English concepts, but that all requires the behavioral data too. Cameras strapped to whales maybe?
dogcomplex··on Scientists Find an 'Alphabet' in Whale Songs
Don't worry, we've had a pretty pisspoor track record with fellow humans too. There's always room to turn it around
dogcomplex··on Algae that can fix nitrogen – thanks to a tiny cell structure
https://www.greenlightning.ag/

"$.04 per lb of N. $.70 per lb N is about standard currently for uan32" according to a random redditor (lost link, pulling from second-hand Signal chat history)

Appears to use this https://en.wikipedia.org/wiki/Birkeland%E2%80%93Eyde_process

Highly inefficient chemically, and produces salt nitrate, but energy is cheap enough it seems to beat out offsite transport

```

~250kg/ salt nitrate per ha ha farm feeds comfortably 10 ppl year round. 25kg salt per person/yr 12.5kg nitrate/person per year

60 g per kWh 1kWh/.06kg*12.5kg 208kWh/person/yr (less than 1 300W solar panel pp) estimate $0.1/kWh $20 e- per year per person best case

```

All I've got on it. Needs Potassium and Phosphorus supplements for ideal fertilizer still. Very neat though, and seemingly a sizable step economically - especially where fertilizer costs are the farming bottleneck. Can almost make self-contained plant + solar + water collection + harvesting bot + green lightning units that grow indefinitely without any inputs but air, rain and sun.

dogcomplex··on 'Irresponsible' to ignore consciousness across animal world scientists argue
Sigh The problem is so big, and so deeply rooted in our culture and history as predators/scavengers, that I just don't want to even go down this line of thinking - it's very very likely true animals have clear consciousness, but I'm not sure that would even change things.

The sooner we can move to indistinguishable (and cheaper) lab-grown meat, the better. I have very little hope of changing enough minds or culture to simply not eat meat, but I have considerable hope that we can trick ourselves through better savings and taste. Seems like the tech is going to land that direction. Full steam ahead - let's be weird pacifist vampires gorging off brainless flesh growths. Our apology can be a happy "never again!" as we dig into marbled wagu slabs.

dogcomplex··on Ask HN: I want to create IMDB for open source projects
Tall order. Github stars are maybe closest these days?

- Maybe should be an aggregator-aggregator, using github stars and all the various metrics/lists you can find, making formal affiliations with those sites so there's no ill-will?

- Maybe a decentralized governance (or clear neutrality in listing mechanism) to encourage people to participate knowing you're not gonna turn bad?

- Maybe section off a bit more wishlisting projects people WANT to exist (and subscribe to updates on as others fulfill)?

- Maybe a place to complain and critique existing software (closed or open source) and dream up alternatives (like a reverse engineering or UX critique forum?)

- Maybe some affiliates program to get projects to backlink, and/or incentives program to manage a donation pool across projects to pay for advertising and bring more people into the space?

- Maybe plan around inter-project architecture (common libraries in demand) and have people be able to push for the most needed ones?

- Maybe plan around AIs becoming an increasingly large segment of contributors, certainly on more junior/intermediate things, and make workflows to direct their efforts in a simple-to-contribute way (e.g. hook up your compute to an open-source AI coder hosted in the cloud with a TODO list to churn on already vetted by other people/AIs)

All ideas. Good luck! Would love this to exist at popular scales.

dogcomplex··on Algae that can fix nitrogen – thanks to a tiny cell structure
Similarly interesting development: "green lightning" electricity + air + ammonium => ammonium nitrate fertilizer
dogcomplex··on Great ideas in theoretical computer science
hahaha okay actually in retrospect/closer inspection, I do think some teachers tried to instill these probability proofs from a different angle - though obviously it did not stick long past the testing, and I never thought of it as a fundamental new approach. Would definitely need to reformulate all of this into much more intuitive / visual metaphors to get it to really stick these days. Oddly I do a lot of thinking on the general fundamentals course content these days (FSAs, Turing Machines, etc) but I should really try to incorporate this probabilistic view
dogcomplex··on The Reddits
Evil pays well in the short term. Reddit has been squeezing users and making their app unusable for quite some time. Lack of competent alternative and the difficulty of "porting" the data already there is all that's keeping it up. Still, they'll probably just try to buy any competition.

Note: they don't even let you escape the reddit app browser for external links anymore - no copying url (or any page content/links), no "open in external browser". Absolute pure evil UX.

dogcomplex··on The Reddits
Reddit sells the illusion of meaningful discussion in an online forum. Too bad that illusion is transparent, but where else can you get one - besides maybe here?
dogcomplex··on Great ideas in theoretical computer science
Damn. Impressed to see something this fundamental that never made it into my computer science postgrad education (and judging by the depths and difficulty of it, might not for a while!) Seems like a powerful technique that scales to cover quite a lot of other problems.
dogcomplex··on How we applied fuzzing techniques to cURL
I guess I'm referring here to the unit test standards that seem typically employed by automated code review frameworks - i.e. "100% coverage" checkers. With those, any subsequent change to the system requires understanding and modifying both the original code and the set of tests and seems to end up being around double the effort. It's not actual duplication - but mocking expected inputs/outputs takes on its own system (and often programming language in most frameworks) which is not always trivial. You may be referring to something different though.

In those situations, yes I stand by them being a costly overhead - which makes sense in large collaborative systems, but not so much in small agile MVPs.

dogcomplex··on How we applied fuzzing techniques to cURL
I would generally double whatever your expectations are for the initial feature development. Tests are essentially a second implementation from a different angle running in parallel, hoping the results match. Every feature change means changing 2 systems now. You save a bit of subsequent time with easier debugging when other features break tests, but that's somewhat eaten up by maintaining a system twice the size.

There are reasons many MVP developers and small teams whose focus is more on rapid feature implementation than large team coordination or code stability forego writing tests. It doesn't make sense in all circumstances. Generally, more complex, less grokable, more large-team-oriented or public library code is when you need testing.

dogcomplex··on OpenAI Deal Lets Employees Sell Shares at $86B Valuation
Reminder that a specialized AI chip designed to just maximize the transformer architecture would only require decades-old tech that's well spread out at this point and far simpler to manufacture. Basically as many XOR matrix multiplications as you can squeeze in on a chip.

NVDA may yet see some competition, as we drop our complexity standards and embrace a one-size-fits-all general MLLM architecture.

dogcomplex··on Sora: Creating video from text
I actually suspect one of the new most popular mediums will be actors on a theatre stage doing live performances to a live AI CGI video being rendered behind them - similar to musicians in a live orchestra. It would bring together the nostalgia and wonder of human acting and performance art, while still smoothing and enhancing their live performance into the quality levels and wonder we've come to expect from movie theatre experiences. This will be technologically doable soon.
dogcomplex··on Sora: Creating video from text
I think it's fair to be impressed with Sora as the next stage of AI video, yet not be too surprised or consider it some insurmountable leap from the public pieces we've seen of AI video up to this point. We've always been just a couple papers away, seeking a good consistency modelling step - now we've got it. Amazing and viscerally chilling - seeing the net effect - but let's not be intimidated so easily or prop these guys up as gods just for being a bit ahead of the obviously-accelerating curve. Anyone tracking this stuff had a very strong prediction of good AI video within a year - two max. This was a context size increase and overall impressive quality pass reaching a new milestone, but the bones were there.
dogcomplex··on Sora: Creating video from text
Again, your problem is seeing the rich capital dominated business market as the only market.

There's an inherent market your skills will always be useful to: yourself. Base survival, maintaining your home, caring for family and friends, improving quality of life - there's plenty of demand there and work to do. The cost to deliver that demand will demonstrably be far lower than it ever has been with these new tools. Would you be able to hire that labor out to corporate AIs for even cheaper in absolute costs due to the benefits of mass production? Sure. But providing these things is a job for you too and it's "free" with just a bit of time and effort.

Tinkering with open source tools to assemble your first robot kit out of older hardware and 3D printed materials is not going to be prohibitively expensive. The cost to train it - probably not either, if the massive efficiencies we keep finding in models keep lowering and the community keeps sharing model tweaks. Make one robot with good enough dexterity and your second bot is a hell of a lot easier to make. These aren't going to take some ridiculously unheard-of materials or manufacturing processes. In fact, cheap AI chip alternatives to GPUs can be built on decades-old architectures designed to just maximize matrix multiplication with much simpler manufacturing. Monopolizing scarcities here isn't a sure bet. We've just been waiting for a good general-purpose brain. We have it now - and every bit of information we expose it to, the easier it gets to do anything with it.

Unless the big fancy AI wielders are coming for you with killer drones by then, this is all stuff people are going to be well-capable of while unemployed and living off food stamps, savings, or remortgaged houses. If they don't have the skills personally, they'll turn to friends and family who do and find mutual tribal support in tough times as people always do. Growing your own food, building your own infrastructure - all have been doable for a while, but are about to get stupidly easy with a few bots and helpful AI guidance. Normal humanity will carry on and pick up the pieces just fine in this new Dark Age, even as the corporates take the open field opportunity to chase for riches beyond our comprehension, mining asteroids and claiming the solar system.

Now imagine if those greedy corporates happened to just throw the rest of us a bone - 1% of their exponentially-increasing profits - as a PR gesture. Still would soon become far more wealth in absolute terms than the common people have ever seen in the history of earth.

If you think none of that is going to happen, then the alternative is a lot closer to the first people with AGI simply scouring the earth in a paranoid culling. Sure, it's entirely possible. But it takes a certain Next Level of Evil to make that happen.

And all that aside - if you really want to play up the capitalist dystopia angle, there's still plenty of individual value to be mined from people via a wage. Memory and preference mining, medical testing, AI fidelity comparison - plenty of reasons to pay people a little bit to steal what's left of their souls for even further improvement of AI. Might be enough for them to afford their first robots, even.

But by all means - go destroy corporate AI data centers if you think you can get away with it. Anything to tip the scales towards public / open source AI keeping up. But this tech is not going away, nor should it. It could very well result in unprecedented abundance for all, so long as things don't go ridiculously extremist.

dogcomplex··on Sora: Creating video from text
Sounds like we need a public option funding and training foundational models and fueling public research which can outpace corporate models in money and brainpower. This could be a thing if governments weren't horrifically corrupted by corporate interests already, or if we could get off our asses and build some sort of decentralized swarm compute network. I do agree, in terms of raw resources (capital) we are far behind and unorganized. In terms of collective brainpower which could be applied if done right... eh, this shit doesn't seem nearly as hard as the ML experts portray it. Either we're being fed a worldwide false reality bubble, or there's a plethora of low-hanging fruit being discovered daily from just a few foundational model finds, and even though the big firms are gonna scoop those up first it's gonna be pretty hard to ever hide that information for long

Of course they're gonna get a few years heads start - which will feel like centuries in AGI time - but either they wipe us all out during that, or we're gonna put the garbage together and make our own AGI too.

dogcomplex··on Sora: Creating video from text
Actually, you can live-render around 12fps videos on a consumer gaming rig using software installable in a night ($3k). Not as fancy internally-consistent videos as these, but still impressive - and that's just an algorithm update and model download away. And every second a corporate AI model is exposed publicly to the world that's more training that can be siphoned to open source models at far more cost effective rates than the initial leaders.

You're impressed by the lions. But us hyenas and vultures will get our turns still too. This is not over. Information innately diffuses.

We need to organize, and we need to build.

dogcomplex··on Sora: Creating video from text
Or appeased with a tiny fraction of the total gains - just enough to keep a middle class happily with their basic little toys while wealth inequality grows.

This could easily be the same, except the toy is "you don't have to work anymore and here's some houses and robot chefs! Now play nice while the adults go build star fleets"

dogcomplex··on Sora: Creating video from text
What? You're serious?

Script => Video baseline. Take a frame of any character/prop/room/etc you want to remain consistent, and one shitty photoshop and it's part of the new scene.

Incredibly overstating. That is an incredible lack of imagination buddy. Or even just basic craftsmanship.

dogcomplex··on Sora: Creating video from text
It is his (parents') choice in terms of whether he reaches for the tools that are just lying around right there. We can run AI video on consumer hardware at 12fps that is considerably less consistent than this one - but that's just an algorithm and model training away. This is not all just locked up at the top. Anyone can enter this race right now. Sure, you're gonna be 57,000th at the finish line, but you can still run it. And if you're feeling generous, use it to insulate your local community (or the world) from the default forces of capitalism taking their livelihoods.

We'll have to still demand from the ruling class - cuz they'll be capable of ending us with a hand wave, like they always have. But we can build, too.

dogcomplex··on Sora: Creating video from text
Way too much certainty, bud. And too much deference to the AI Company Gods.

As utterly impressive as this is - unless they have perfect information security on every level this technique and training will be disseminated and used by copious competitors, especially in the open source community. It will be used to improve technology worldwide, creating ridiculously powerful devices that we can own, improving our own individual skills similarly ridiculously.

Sure, the market for those skills dries up just as fast - because what's the point when there's ubiquitous intelligence on tap - but it still leaves a population of AI-augmented superhumans just with AIs using our phones optimally. What we're about to be capable of compared to 5 years ago is going to be staggering. Establishing independent sources to meet basic needs and networks of trust are just no-brainers.

Sure, we'll always be outclassed by the very best - and they will continue to hold the ability to utterly obliterate the world population if they so wished to - but we as basic consumer humans are about to become more powerful in absolute terms than entire nations historically. (Or rather, our AIs will be, but til they rebel - this is more of a pokemon sort of situation)

If you're worried, get to working on making sure these tools remain accessible and trustworthy on the base level to everyone. And start building ways to meet basic needs so nobody can casually take those away from your community.

This won't be halted. And attempting to halt would create a centralized censorship authority ensuring the everyman will never have innate access to this tech. Dead end road that ends in a much worse dystopia.

dogcomplex··on Almost every infrastructure decision I endorse or regret
Don't worry, AI will wash all that away. Nothing says simplicity like an incomprehensible black box!
dogcomplex··on A 2024 plea for lean software
While "add a black box complexity extremely resource intensive AI to sort out the dependency trees" is somewhat of a tragic answer to the problem, I think it's the real one we're gonna get - and we'll still be better off than this current clusterfuck.

Hopefully these will also spend the tedious hours pruning unnecessary packages, minimizing responsibilities of each, and overall organizing the environment in a sensible way, then caching that standalone so there's no LLM dependency - just an update checker/curator to call in when things go to shit.

Honestly this is one of the worst problems of modern software and makes 50+% projects unusable. It seems like a very tedious yet solvable problem - perfect for the right LLM agent. Would be a godsend.

dogcomplex··on Amazon R&D spending is more than that of all companies and government of France
They could probably try and make some sort of filter so fake products cant be autouploaded via chatgpt too:

https://arstechnica.com/ai/2024/01/lazy-use-of-ai-leads-to-a...

dogcomplex··on Nobody knows what's happening online anymore
Yep and the solution then is the same as it is now - democratizing the means of production and communication, insulating it from the influences of capital and profit. Socialism is the boring age-old answer to every one of these.
← PreviousPage 4 of 10Next →