Tools like Claude Code and OpenAI’s Codex CLI have boosted my productivity massively. They already handle about 90% of the coding work, and I just step in to finish the last 10%. Every month they get better, maybe in a year it’s 95%, in two years 97%, in three years 98%. We can all see where this is going.
You're an early adopter, so you're seeing massive gains. But eventually everyone gets the same productivity boost and that becomes the new baseline expectation.
Any clever prompting techniques that give you an edge today will evaporate quickly. People figure out the tricks, models absorb them, and tools automate them away.
There's no substitute for actually coding to learn software development. For new engineers, I'd strongly recommend limiting AI code generation on real work. Use it to explain concepts, not do the work for you. Otherwise you'll never develop the judgment to know what that 10% actually is.
You are fortunate to have the pre-AI background experience to know how to recognize the 10%. People graduating college right now may never be so fortunate.
This doesn't just apply to software development - it applies across the board in any form of knowledge.
An example would be economics - an LLM can spit out a bunch of words about an economic model. But if you don't take the time to learn, visualise and understand it for yourself - it means nothing to you. And in that case, if you already possess mastery why would you waste your resources playing around with an inferior tool to you?
Absolutely agreed. Thinking anything else is nothing but cope, and these comments are FULL of it. it would be laughable if they weren't so gate keepy and disengenuous about it.
two years ago, I opened a PR asking for an LLM commit feature, and they flat-out said they weren’t doing it. Meanwhile, Cursor was eating their lunch and lapping them twice. I couldn’t believe how complacent and out-of-touch they were—it was pure laziness dressed up as “product focus.” And let’s not forget the ancient bugs rotting in their backlog that they refuse to fix. It’s like they actively don’t care about their users.
Claude Code has honestly made me at least 10x more productive. I’ve burned through about 3 billion tokens and have been consistently merging 5+ PRs a day, tackling tons of tech debt, improving GitHub Actions, and making crazy progress on product work
only 10x? I'm at least 100x as productive. I only type at a measly 100wpm, whereas Claude can output 100+ tokens a second
I'm outputting a PR every 6 minutes. The reviewers are using Claude to review everything. It used to take a day to add 100 lines to the codebase.. now I can add 100 lines in one prompt
If I want even more productivity (at risk of making the rest of my team look slow) I can tell Claude to output double the lines and ship it off for review. My performance metrics are incredible
So no human reads the actual code that you push to production? Are you not worried about security risks, spaghetti code, and other issues? Or does Claude magically make all of those concerns go away?
This is only the beginning. I can see myself having 100 Claude tasks running concurrently - the only problem is edits clash between files. I'm working on having Claude solve this by giving each instance its own repo to work with, then I ask the final Claude to mash it all together as best it can
What's 100x productivity multiplied by 100 instances of Claude? 10,000x productivity
Now to be fair and a bit more realistic it's not actually 10000x because it takes longer to push the PR because the file sizes are so big. Let's call it 9800x. That's still a sizable improvement
I also have this feeling that I'm 2-10x more productive. But isn't it curious how a lot of devs feel this way, but no devs that I know have the experience that any of their colleagues have become 2-10x more productive?
<raises hand> Our automated test folks were chronically behind, struggling to keep up with feature development. I got the two assigned to the team that was the most behind set up with Claude Code. Six weeks later they are fully caught up, expanding coverage, and integrating AI code review into our build pipeline.
It's not 10x, but those guys do seem like they've hit somewhere around 2x improvement overall.
Sometimes 10x can mean that I start things that I would have never started before, knowing it would take a long time. Or that I can have any of the agentic stuff "explore" libs, stacks and frameworks I wanted to look at, but had no time. Or distill some vague docs and blog posts to find common use cases for tech x. And so on.
It's not always a literal 10x time for taskA w/ AI vs taskA w/o AI...
What type of work do you do and what type of code do you produce?
Because I've found it to work pretty amazingly for things that don't need to be exact (like data modeling) or don't have any security implications (public apps). But for everything else I end up having to find all the little bugs by reading the code line by line, which is much slower than just writing the code in the first place.
How do you maintain high confidence in the code it generates ?
My current bottleneck is having to review the huge amounts of code that these models spit out. I do TDD, use auto-linting and type-checking.... but the model makes insidious changes that are only visible on deep inspection.
You have to review your code for quality and bugs and errors now just as you did last month or last year. Did you never write bugs accidentally before?
We're all bottlenecked on reviewing now. That's a good thing.
There was a greater awareness of exactly what I'd written. By definition, I would not have written those bugs in, as long as I had known edge cases in my mind.
Lapses of judgement and syntax errors happen, but they're easier to spot because you know exactly what you're looking at. When code is written by a model, I have to review it 3 times.
1st to understand the code. 2nd to identify lapses in suspicious areas. 3rd to confirm my suspicions through interactive tests, because the model can use patterns I'm unfamiliar with, and it takes me some googling to confirm if certain patterns used by the model are outright bugs or not. The biggest time sink is fixing an identified bug, because now you're doing it in someone-else's (model's) legacy code rather than a greenfield feature implementation.
It's a big productivity bump. But, if reviewing is the bottleneck, then that upper bounds the productivity gains at ~4x for me. Still incredible technology, but the death of software-engineering that it is claimed to be.
Honestly, at this point, I'll just work 2 hours a day since I'm already producing 5-10x the output of my coworkers who aren't leveraging Claude Code at the medium-sized startup where I work. It really makes you wonder what's holding people back if they haven't at least doubled their productivity by now
Peaked? Nah, it's barely started. Wait till we get decent SWE agents reliably writing good code, probably later this year or next. Once AI moves beyond simple boilerplate, the productivity boost will be huge. Too soon to call hype when we've barely scratched the surface.
I asked copilot to write me a Typescript function today
I had two defined types, both with the exact same field names. The only difference is one has field names written in snake_case, and the other had names written in camelCase. Otherwise the exact same
I wanted a function that would take an object of the snake_case type, and output an object of the camelCase type. The object only had about 10 fields
It missed about half of the fields, and inserted fields that didn't even exist on either object
You cannot convince me that AI is anywhere near to this level if it cannot even generate a function that can convert "is_enabled" to "isEnabled" inside an object
Every time I try this stuff I'm so disappointed with it. It makes me think anyone who is hyped about it is an absolute fraud that does not know at all what they are doing
This mirrors my experience, and I've tried nearly every popular model available, numerous times over the last year or so. I'm having trouble understanding where all the hype is coming from -- is it literally just all marketing? I refuse to believe any half-decent SWE can be bullish on AI-for-code after novelty of the first week wears off.
You get out what you put in. Of course if you provide one sentence of context(and some implicit context from the environment) you aren't going to get magical result.
The test was "can I get it to generate this while spending less effort than it would take me to just write it" and it failed miserably. And this was a super low effort, small boilerplate problem to solve. This is the sort of problem it has to solve to be remotely useful
If it cannot do that, then why is anyone saying it is a productivity booster?
My response could only possibly be that I haven't had that issue. I've asked for relatively complex changes to codebases(mainly python), and had very little in the way of trouble.
The more code you ask it to generate, the higher the chances that it will introduce an issue. Even if the code compiles, subtle bugs can easily creep in. Not unlike a human programmer, you might say, but a human programmer wouldn't hallucinate APIs. LLMs make entirely unique types of errors, and do so confidently.
You really need to carefully review every single line, which sometimes takes more effort than just writing it yourself. I would be particularly wary of generated code for a dynamic language like Python.
Honestly if long context (that doesn't start to degrade quickly) is what you're after, I would use Grok 3 (not sure when the api version releases though). Over the last week or so I've had a massive thread of conversation with it that started with plenty of my project's relevant code (as in couple hundred lines), and several days later, after like 20 question-aswer blocks, you ask it something and it aswers "since you're doing that this way, and you said you want x, y and z, here are your options blabla"... It's like thinking Gemini but better. Also, unlike Gemini (and others) it seems to have a much more recent data cutoff. Try asking about some language feature / library / framework that has been released recently (say 3 months ago) and most of the models shit the bed, use older versions of the thing or just start to imitate what the code might look like. For example try asking Gemini if it can generate Tailwind 4 code, it will tell you that it's training cutoff is like October or something and Tailwind 4 "isn't released yet" and that it can try to imitate what the code might look like. Uhhhhhh, thanks I guess??
