In 2019, this is sci-fi
In 2023, it can help engineers understand the complexity of taxes, translate some tax rules into code, help refine edge cases and fix mistakes.
In 2030, ???
They can’t imagine we could also just hit a plateau and have no progress for a decade until some other major breakthrough happens.
AI art is hitting a plateau or has hit a plateau. Progress came quickly for months but nowadays each update is just a bit better details with more resolution and better hands. That’s it.
If you have AI anxiety, just read my comment history.
> AI art is hitting a plateau or has hit a plateau.
I am not sure I am following. Are you saying AI art has hit a plateau since stable diffusion and dall-e 2 got released? This is less than a year ago. Look for plateau at the scale of a 5 to 10 years, not 6 months.
PLs benefit from being targeted at a specific domain dramatically that NL just can't.
It was more an indexing question and a factor of me remembering the concept of a theater play, but nothing else.
I was really impressed when the LLM found the answer right away. Until I realized it was a lie.
I had to made it print a list of 50 plays, and I found what I was looking for.
I realize it’s not a good usage for a LLM but I think it’s how it’s gonna be with those
Those are called programming languages, we have quite a lot of them.
Help engineers understand taxes: it is pretty much part of the GPT-4 demo. You can copy the tax code in it and ask questions about it. As a matter of facts, it helped me with some taxes questions I had this year, although it was ChatGPT and I didn't paste the tax document in it.
Translate some tax rules into code? I did something similar last week for a complex billing project I am working on. It didn't give me the best variable names and made up a few some functions but the formulas were correct and I reused a good chunk of it.
Fixing the logic of the code it just wrote? I think there are enough examples of that on Twitter.
Doesn't it still have problems with outright hallucinating a bunch of crap?
If so, the dangers of using something flaky/unreliable for potentially life changing decisions...
But whenever I have to deal with a complex question such as taxes, I often find it very useful to break down a problem, rephrase and provide context around my question and offer different leads to follow or cross reference.
I see it as GPT spiting out a skeleton of an answer onto which I can attach some meat. It is very useful when I don't really know the domain and therefore how to attack a problem.
And I'm not even sure any of that is a bad thing. How do we deal with the idea that the AI industry might create a benevolent God?
Depending on the quarter only 10% to 25% of my time is writing code. The other is dealing with: finding high level reliability concerns, figuring out what solutions to implement, diving into complex acute production issues, evaluating trade-offs for fixes, finding ways to make things more efficient, negotiating what to prioritize with other teams/orgs/leadership, etc.
Could an AI do some of these things? Sure. Are they going to be able to figure out the right stakeholder to talk to, negotiate with them, get buy-in and decide on overall designs/solutions? No.
I'd be a bit concerned if I was a frontend developer (though any decent one could pivot to something else.) however.
I guess what I mean to say as SRE/Systems, we code so little already that it being able to spit out a CF template wouldn't reduce the needs for cross-functional ops type work.
I so wish you had not done that: the number 25 is now tainted in my workmates' collective memory, no one of them is willing to touch it, yet I fear we shall need it again soon.
You should have instead "pulled that number out of a hat":
https://idioms.thefreedictionary.com/pull+out+of+a+hat
As you can see by the following, there is a difference:
https://idioms.thefreedictionary.com/pull+out+of+ass
Perhaps you can be forgiven this once, being a back-end developer, but, in the future, please consider your audience.
or so it goes, no idea if this is going to actually happen.
It already has. Every single one of my coworkers is using Copilot. If you mean in terms of the job market, the shift to more declarative supervision of the tool rather than imperative implementation of the work yourself means that the bar for skill is far lower.
There will come a time when product managers and UX designers just use AI to generate Figma mocks, and AI generates an entire scaleable platform from just that. Will they need developers? Sure, but maybe 1. Certainly not 8-10.
And all of you still got a job. Copilot has been out for a time now, if it was transformative we would see companies releasing a ton of products and fix all the bugs and then fire most of their developers etc the past half year, but that didn't happen.
Making developers a bit more productive wont have an impact on developer jobs. It changes your day to day job a bit, but that is like switching to another IDE, it wont transform the industry. Even if it reduces the need for developers by an entire 10% it is not that big of a deal, it would be worth many billions but for an average developer they wont notice much of an impact. It is like the productivity improvement of giving developers another screen, a better computer, a better IDE etc, those are nice but they didn't transform the industry.
Seems a little stupid if your competitor has way more productivity?
This has never saved anyone's job from automation. Ask the coal mining industry in the US. In fact, capitalism will do the opposite because investors own most companies and investors like to see reductions in head counts.
> What if their competitor decides they're going to retain all of their engineers and produce 10x more than they did before?
There are very few companies where "more code" = "more profit". Code is a cost center and support function of the business. Look at all the companies that recently fired 20%+ of their coders and are more profitable than they were before.
If your coders are 10x as productive and your business is growing in a healthy way, your bottleneck is going to be sales and marketing. It probably already was sales and marketing. You're not going to retain all the people you don't need.
The alternative would be built to help users leverage AI to do their taxes. Upload your docs and it’ll get you 90% of the way there or something like that. AI will help developers build that alternative faster. Almost every piece of software will need to be reimagined from the ground up in this way. That is the job security