421–430 of 1001 posts
Earlier quoted context omitted.
The problem is that we haven't invented AI that reads the crap that other AIs produce - so the burden is now on the reader to make sense of whatever other people lazily generate.
But we do. The same AI that generates can read and reduce/summarize/evaluate.
I don't want to read work that someone else couldn't be bothered to write.
Earlier quoted context omitted.
I agree with most of your fears. There is one silver lining, I think, about superintelligence: we always thought of intelligent machines as cold calculators, maybe based on some type of logic symbolic AI. What we got instead are language machines that are made of the totality of human experience. These artificial intelligences know the world through our eyes. They are trained to understand our thinking and our feelin…
This gives me a little hope.
Earlier quoted context omitted.
> marketing fluff for the App Store If it’s fluff, why do you put it there? As an App Store user, I‘m not interested in reading marketing fluff.
Because it’s required? I’ve released over 20 apps, over the years, and have learned to add some basic stuff to each app. Truth be told, it was really sort of a self-deprecating joke. I’m not a marketer, so I don’t have the training to write the kind of stuff users expect on the Store, and could use all the help I can get Over the years, I’ve learned that owning my limitations, can be even more important, than knowing…
I'm not tired, I'm afraid. First, I'm afraid of technological unemployment. In the past, automation meant that workers could move into non-automated jobs, if they were skilled enough. But superhuman AI seems now only few years away. It will be our last invention, it will mean total automation. There will be hardly any, if any, jobs left only a human can do. Many countries will likely move away from a job-based market…
Behind either the fear or the hope, is necessarily some utter faith that a certain kind of future will happen. And I think thats the most interesting thing.
Because here is the thing, in this particular case you are afraid something inhuman will take control, will assert its meta-Darwinian power on humanity, leaving you and all of us totally at their whim. But how is this situation already not the case? Do look upon the earth right now and see something like benefits of autonomy or agency? Do you feel like you have power right now that will be taken away? Do you think the mechanism of statecraft and economy are somehow more "in our control" now then when the bad robot comes?
Does it not, when you lay it out, all feel kind of religious? Like that its a source, driver of the various ways you are thinking and going about your life, underlayed by a kernel of conviction we can at this point only call faith (faith in Moores law, faith that the planet wont burn up before, faith that consciousness is the kind of thing that can be stuffed in a GPU). Perhaps just a strong family resemblance? You've got an eschatology, various scavenged philosophies of the self and community, a certain but unknowable future time...
Just to say, take a page from Nietzsche. Don't be afraid of the gods, we killed them once, we can again!
So what's left for humans?
We very likely won't have as many human software testers or software engineers. We'll have even fewer lawyers and other "credentialed" knowledge worker desk jockeys.
Software built by humans entails humans writing code that has not already been written -- by writing a lot of code that probably has already been written and "linking" it together, etc. When's the last time most of us wrote a truly novel algorithm?
In the AI powered future, software will be built by humans herding AIs to build it. The AIs will do more of the busy work and the humans will guide the process. Then better AIs will be more helpful at guiding the process, etc.
Eventually, the thing that will be rewarded is truly novel ideas and truly innovative thinking.
AIs will make varioius types of innovative thinking less valuable and various types more valuable, just like any technology has done.
In the past, humans spent most of their brain power trying to obtain their next meal. It's very cynical to think that AI removing busy work will somehow leave humans with nothing meaningful to do, no purpose. Surely it will unlock the best of human potential once we don't have to use our brains to do repetitive and highly pattern-driven tasks just to put food on the table.
When is the last time any of us paid a laywer to do something truly novel? They dig up boilerplate verbiage, follow standard processes, rinse, repeat, all for $500+ per hour.
Right now we have "manual work" and "knowledge work", broadly speaking, and both emphasize something that is being produced by the worker (a construction project, a strategic analysis, a legal contract, a diagnosis, a codebase, etc.)
With AI, workers will be more responsible for outcomes and less rewarded for simply following a procedure that an LLM can do. We hire architects with visual / spatial design skills rather than asking a contractor to just create a living space with a certain amount of square feet. The emphasis in software will be less on the writing of the code and more on the impact of the code.
I am tired and hungry… The thing I’m tired of is elites stealing everything under the sun to feed these models. So funny that copyright is important when it protects elites but not when a billion thefts are committed by LLM folks. Poor incentives for creators to create stuff if it just gets stolen and replicated by AI. I’m hungry for more lawsuits. The biggest theft in human history by these gang of thieves should be…
Lawsuits based on what? Copyright? People crying for copyright in the context of AI training don't understand what copyright is, how it works and when it applies. What they think how copyright works: When you take someones work as inspiration then everything you produce form that counts as derivative work. How copyright actually works: The input is irrelevant, only the output matters. Thus derivative work is what exp…
Now if the likes of Openai and Google keep the model weights private and just provide generated text, they can try to filter for derivative works, but I don't see a solution that doesn't leak. If a model can be coaxed into producing a derivative work that escapes the filter, then boom, unlicensed copy was provided. If I tell the model to mix two texts word by word, what filter could catch this? What if I tell the model to use a numerical encoding scheme? Or to translate into another language? For example assuming the model knows a bunch of NYT articles by heart, as was already demonstrated: If have it translate one of those articles to French for me, that's still an unlicensed copy!
I can see how they will try to get these violations legalized like the DMCA safe-harbored things, but at the moment they are the ones generating the unlicensed versions and publishing them when prompted to do so.
I am tired and hungry… The thing I’m tired of is elites stealing everything under the sun to feed these models. So funny that copyright is important when it protects elites but not when a billion thefts are committed by LLM folks. Poor incentives for creators to create stuff if it just gets stolen and replicated by AI. I’m hungry for more lawsuits. The biggest theft in human history by these gang of thieves should be…
Earlier quoted context omitted.
I would never have imagined hackers becoming copyright zealots advocating for lawsuits. I must be getting old but I still remember the Pirate Bay trial as if it was yesterday.
I'm not sure if you're being disingenuous, or if you genuinely don't understand the difference. Pirate Bay: largely facilitating the theft of material from large corporations by normal people, for generally personal use. LLM training: theft of material from literally _everyone_, for the purposes of corporate profit (or, well, heh, intended profit; of course all LLM-based enterprises are currently massively loss-makin…
> But I’m pretty sure I can do without all that ... test cases ... Test cases? I did a Show HN [1] a couple of days back for a UI library built almost entirely with AI. Gpt-o1 generated these test cases for me: https://github.com/webjsx/webjsx/tree/main/src/test - in minutes instead of days. The quality of test cases are comparable to what a human would produce. 75% of the code I've written in the last one year has b…
I am tired and hungry… The thing I’m tired of is elites stealing everything under the sun to feed these models. So funny that copyright is important when it protects elites but not when a billion thefts are committed by LLM folks. Poor incentives for creators to create stuff if it just gets stolen and replicated by AI. I’m hungry for more lawsuits. The biggest theft in human history by these gang of thieves should be…