[flagged]
That said, I think ML advances may possibly usher in the next gen of low-code tools that may liberate a large portion of web devs from being human LLMs.
I think you are right w.r.t. Frontend, it's significantly easier than regular programming as you can(will be able to) simply enumerate a list or tree of the various style choices, generate them all, and allow people to A/B/C/D/...ZYX test between them.
Honestly, we shouldn't need AI/ML for this. We have so many well established need and use patterns that we should already have standardized "lego" style functionality blocks, optionally with interchangeable wrappers to provide customization options (common automatic options or even programmable ones).
The reason we're still all reinvinting wheels all the time is, frankly, because it's too easy to do our own thing. I would never advocate for this (and I'm not an MS guy), but perhaps if Microsoft had won the world and Linux had never existed, we would all be using the same MS libraries and possibly be further down the road than we are.
What I've observed over 20 years is the opposite - a continual expansion, even an explosion of competing (often uncompelling) alternatives.
What I am imagining is LLM crafts the code in such a way that you can look at a small set of certain files and get a real understanding of the story of that program or service. Some people write code like that but most (including me) can't unless I spend a lot of time crafting, which is not a luxury we get at work.
Pedagogical code is probably the word I am looking for!
With that you could get an LLM to write a new cryptography library, and assuming you are a crypto expert, you could quickly verify that it is doing the right thing, as it makes it obvious.
The day one of these find new algorithms to solve problems with better complexity or simpler code that state of the art, I'll wake up. When I give a LLM a computational geometry problem, it's exactly like a student trying to bullshit his/her way through an exam without any actual deep understanding.
For example, I ask for an algorithm to compute Laguerre Voronoi diagrams (usually not available in books or code examples), and I get answers for plain Voronoi diagrams, because it's what you will find in many books and code samples. Generating boring but necessary code, in moderation, is a win.
IIRC there are some people looking at theorem proving with LLMs, but the reality is they don't have to do anything foundational or groundbreaking to be of value in assisting, supplanting, or replacing the vast majority of people who interact with computer code.
We are talking about the field that openly just memorized leetcode questions to "make it", right?
Why should we have such high standards for tools when we don't even apply them to the "best and brightest"
They’re just people. Not divine mandates. There is zero real obligation to serve contemporary socio-political and economic norms.
Not really finding any of this progress in computing shocking though. Unix and that model wastes a lot of resources dealing with strings.
What a shock we reduce the amount of fluff to just enough symbolic logic to instigate appropriate electron state to solve a problem, rather than brag about a new DSL to tokenize, parse and logic to process bespoke template syntax in yet more asinine “file formats” …once you remove the chimps chasing Shakespeare, superfluous state doing nothing but propping up venture backed coder boot camp grads jobs we find super powerful software.
Shocking, I’m shocked.
Huh?
Disagree with the use of that term.
If anything copyright is the heist. Disney subsists on back breaking labor of others while it wags around papers saying it owns things that make it rich enough to avoid real work.