Don't get discouraged too soon.
There's still a long way to go until we reach 'models do everything and make everyone irrelevant'.
Yes, there will eventually be just one application which can 'do everything', but somebody has to build that.
Humans must still review the generated code and figure out the 'business logic' - eg. what they (humans) want from the app, even though the code typing will be largely automated.
Also, for the AI to develop to those levels, society and the economy as a whole needs to keep being relatively stable, which is not guaranteed, given challenges like political instability, possibility of war, climate..
But yeah, things are changing fast for us devs (and many more professions out there), AI makes solving problems easier and cheaper, so demand for problem solvers will inevitably go down.
On the bright side, those same people who are not in demand also have access to these tools, so they can use it to create their own thing with less effort.
As for getting old, yeah, I think about this more than I used to.
I dedicate a big amount of time to staying in shape and doing physical activities, going into nature, using my body.
I feel like I've wasted decades of beauty while typing all that code and not being in nature.
Eg. The external text is fed into a LLM that doesn't have access to any tools.
The layer is instructed to produce intermediary output, which is then fed to the second layer.
Second layer also without tools, instructed to further validate the output of the first layer and produce structured output for the third (or nth) layer.
The third (or nth) layer has access to tools and ability to perform side effects.
> I genuinely believe that C++ is just a toxic wasteland of time. It's not even that productive, fun to work with or secure. I would take C, Rust, Golang, Python anywhere, anytime.
Please reconsider bashing it that much, if not out of respect for the people who use this tool daily, then out of the observation that the program you used to write this message was probably written in or uses libraries written in C++.
It's a powerful tool and has been used to create amazing things in competent hands, so maybe its flaws, that we like to bash so much, are a great thing - it's what drove people to create all this other languages that we love to experiment with.
It's still semi-compatible with C, templates are about as close to macros as you'll get with a syntax; and the algorithms parts of stdlib are in many ways brilliant, thanks to Stepanov.
C++ isn't a person, and so we're not hurting its feelings by pointing out that it's terrible. If we hurt Bjarne's feelings by criticising his baby I'm sure he'll get over it.
I don't see much difference between pointing out that C++ is a bad idea despite the fact that some good software was written using C++ and pointing out that slavery is a bad idea despite the fact that some famous US landmarks were built using slave labour.
You know, you can replace "C++" in your comment with "PHP" and it would still be just as correct. But that doesn't make PHP (or C++) any better than it actually is.
Also, "reconsider bashing it that much, if not out of respect for the people who use this tool daily", wow. Surely you don't mean that respecting people who are in a tough situation (e.g. having to write C++ daily) involves never pointing out that the situation they are in is, indeed, tough and could be better?
Funny how in making this prediction, the author of gmail seems to forget that google also has.. gmail and a gazillion other products which bring in revenue.
ChatGPT can't find videos or images or company home pages or the tax office website or the electrical utilities, etc.
It could, but that would turn it into a search engine and I think it's reasonable to assume that Google would come up with its own version of ChatGPT sooner than OpenAI would build a search engine.
Personally, I don't want to waste my energy reading generated text.
At the core of it, people write in order to transmit some deeply distilled messages about life. It is about sharing the experience of being alive, either as advice or warning about what might happen or will inevitably happen to us.
The love, the pain, the emotions, the fear of death, the acceptance - good writing is where we read these things in between the lines, where we feel and empathize with the author and as a result gain some deeper insight which helps us adapt to the ever changing circumstances around us.
There's a lot more encoded in that text than just semantic meaning of words or phrases.
Not that all human writers have the 'talent' to encode more than that, but the ones who do manage to shift something inside us.
As of right now, I can feel that a text was generated. Same with images - and with sound. I can't exactly explain it, but it's the same kind of 'plastic' feeling and it's similar regardless of the form (text, image, sound).
I'd really like us to be able to keep this edge over the algorithms, but this might be impossible in the long run.
Most of the top search results for such keywords are usually written by $30/article writers anyway. None of them have real subject matter expertise.
An AI that aggregates all the knowledge from the internet is more likely to be correct than some writer you hired off UpWork to pump out 100 keyword rich articles on 30 different topics.
>At the core of it, people write in order to transmit some deeply distilled messages about life.
People write for all kinds of reasons, including this one. Agreed that this case isn't reproducible by machines...but the other 50-80% of writing, including technical writing, might be.
It's writing correct Clojure code and Clojurescript components, it's producing the CSS and the SQL migrations and it has context of what it is working on, so you can just say "repeat the same CSS but use a green tint".
I was also trying Clojure code. Quicksort worked, radix sort and levenshtein / edit distance did not work. I also tried asking it to make code to visualize the mandelbrot set in ascii which someone else apparently did in Erlang and only had to make a few changes. It tried to use clojure.math.complex which does not exist.
However you turn it around, you are the Universe experiencing itself subjectively. That's easy to understand without any drugs.
Having felt that during your trip might have led you to the wrong conclusion that your social life will somehow magically be improved by this knowledge alone.
It could be, but not without a lot of good old work and effort.
Even after all of that, there's no guarantee that you'll get what you want and won't get what you don't want.
A lot of people waste their lives with or without the drugs, so maybe it wasn't just this substance that led to those setbacks.
- Yes, doing this at a meaningful scale will require large bioreactors with a price tag of tens to hundreds of millions of dollars. You will need LOTS of those reactors to get production beyond 1% of the world's current coffee consumption.
- With all of these, one has to ask what the point is if the bioreactor inputs are sugar (from plants) and other nutrients. If you want sustainability benefits for X where X = plant-based product, you're probably better off improving the current agricultural practices of growing X rather than growing X in a bioreactor.
- That's one of the loftier visions in biotech: decentralized manufacturing of basically anything. Anything along the lines of a convenient personal bioreactor currently seems like it might as well be a century away or more, but I'd certainly be an early adopter:)
I would guess that it depends on the final product. This works for coffee because it gets grinded to a fine powder. Cocoa might work as well, tobacco for e-smokig, etc.
Yes, there will eventually be just one application which can 'do everything', but somebody has to build that.
Humans must still review the generated code and figure out the 'business logic' - eg. what they (humans) want from the app, even though the code typing will be largely automated.
Also, for the AI to develop to those levels, society and the economy as a whole needs to keep being relatively stable, which is not guaranteed, given challenges like political instability, possibility of war, climate..
But yeah, things are changing fast for us devs (and many more professions out there), AI makes solving problems easier and cheaper, so demand for problem solvers will inevitably go down. On the bright side, those same people who are not in demand also have access to these tools, so they can use it to create their own thing with less effort.
As for getting old, yeah, I think about this more than I used to.
I dedicate a big amount of time to staying in shape and doing physical activities, going into nature, using my body.
I feel like I've wasted decades of beauty while typing all that code and not being in nature.