Boris is right. "Coding" is solved.

Hard disagree, till llms can predict a better architecture than my own, at 6y of xp, then we'll talk. Me having to constantly babysit Opus 5, is not solved

Define "coding"

If it's just typing code, then it's solved, but if it's a synonym to programming, software engineering, etc, like I regard it to be, then it's not solved. Gemini says this for example, it's how I think about coding too. Webster is more vague too, but I feel line you can put lots of processes as part of coding, like thinking more higher level, abstracting correctly (and not overengineering shit), etc etc Basically, not just typing code.

When that happens, I'll be the happiest of all. I like AI. I use it regularly. I wouldn't consider myself a senior engineer or anything like that yet, but for someone so early in their career to have to steer it and correct it so much is meh tbh. Each breakthrough is a bit dissappinting, specifically opus 5, I really feel like I was less frustrated / less babysitting 4.5.

No. He isn't. And you prove you don't code.

Define "coding"

I completely agree with all of your takes here based on my own experience of closing agentic loops all over As a result, I can have those architectural conversations, build the right primitives, and then give my agents loose prompts and they get it. 💯

They created a solution to create a new problem so they can charge more to solve the new problem the new solution created

Hmmm, you need to define "Coding".

Conversations like this is why the other engineering disciplines make fun of us.

Please define "coding" 🤔

If coding really is 'solved', then the interesting part moves one step down the pipeline. Can I trust what it wrote? Can I simplify the crap it introduced? Can I break it before production does? I’ve basically started treating AI-generated code as guilty until proven innocent 😂 Been building separate swarms for simplification + adversarial review for exactly this reason. Generation is becoming the cheap bit.

Aviation is solved, landing is not yet solved. Fix incoming.

The implication of saying coding is solved is that you have little to no bugs. Splitting bugs as an outcome from coding is the worst kind of semantics.

that's cool but what about the bugs

Coding isn’t solved. Typing code is. The hard part is deciding what should change, verifying it, coordinating parallel work, and making sure five correct changes still form one correct system.

Boris is a great at sophistry based marketing. He throws out vague things that sound like maybe something but on further inspection are essentially meaningless.

If your definition of "coding" is just getting a hello world app to run, yes it's solved. If it's anything more complicated than that, it's not solved - because otherwise you can just argue that every program is just a hello world app but the lack of any additional requirements or features are just 'implementation bugs'.

Center my btn btch

Indeed. Great video!

Fully agree with your video !

"coding is solved, bugs are not" is doing heavy lifting there. the bug part was always the job

AI is solving typing code, when was this ever the bottleneck to begin with?

It was solved last year and yet here we are 🤷‍♂️

I don't know how this is up for debate. I have 20 years of coding experience. I do not touch a single line of code now. If this doesn't define 'solved', I don't know what does. What's left is coming up with ideas, strategizing, orchestrating, & validating results match ideas & work.

Spelllinkg iz sooolved. Spelllinkg misteiks are naht sooolved. Meiks sens 😛

请关注一下昨天 3h40m 的 claude 宕机，官方整体装死

Define “coding”

They only tested the happy flows

At this point you cannot justify humans writing code for anyone outside plant control (and similaire). It's already becoming a massive security concern, as we are forced to raise standards humans become a liability. this is goig to apply to a hell of a lot of high paying jobs

Hallucinations

It's solved thanks to 5.6 sol and codex.

awesome video Theo!

i don’t think people claiming this know any history about human knowledge. this is the kind of response we’ve seen over and over. one of the most famous cases being the one in which physics was claimed to be solved, already. just few years later, quantum physics was born. whenever i see that some xyz is solved, i always remember the story of Michelson or Jolly, ensuring that all physics was already known, it was just a matter of precision. Jolly even advised Max Planck to avoid entering the field of theoretical physics because it was already done. thankfully, Planck ignored this advise, and became the main contributor of the field that allows us today to be using our mobiles

boris is wrong, opus 5 doesnt understand what a human hand is, or how many hands a human is supposed to have, and suddenly wants to tell me about ireland household light switches and just takes forever to decode Amandas safety bible and doesnt get shit done/..;4r

Nope, it isn't solved... yet. ;)

principal engineer btw performative

I like how Theo has his niche definition of “coding” and pushing that to people.

qpal ka ba ya?