aerhardt

For all his weirdness and moral failings, I don’t see Altman saying things like whites being under apartheid in the US. And worse. Multiple times a day. Every day.

Dario Amodei said in the most recent interview with Dwarkesh that Anthropic currently gets achieves an increase of around 20-30% coding productivity, which tracks with my experience. What do you do to reap orders of magnitude more?

Also, how much more money do you make? Or are you working less?

I've been selling and managing my own projects for over six years, so I don't count myself in the antisocial camp. But LLMs haven't changed the fact that I like to have at least five hours of deep work every day.

I hope you appreciate the irony of saying that in a thread where we are discussing that OpenAI's main competitor is engaging in blatantly anti-consumer behavior.

You're at least describing someone who sounds hard-working... what's the problem?

I'd be more concerned if I was someone who signed up to play ping pong two hours a day and do a bi-weekly commit.

There was a time not so long ago where I was watching "a day in the life of a software engineer" videos on Youtube and I was wondering if some of these were parodies. I still remember one in particular which I'm pretty sure was a parody, but it was only marginally distinguishable from the others.

They're worried about the next step, when we haven't quite digested the current step?

I'm actually surprised that at the scale that AI is being used, we haven't seen more of this - or worse.

Of course a counterfactual is needed, absent clear separation of causes and links to effects, neither of which the productivity metrics on their own establish. This is also widely known and talked about in econ circles in the face of this very data.

> "information technology" generally didn't increase productivity

Do you think it'd be viable to run most businesses on pen and paper? I'll give you email and being able to consume informational websites - rest is pen and paper.

When they decide to touch something as they go, they often don't improve it. Not what I would call "refactoring" but rather a yank of the slot machine's arm.

I'm building a website in Astro and today I've been scaffolding localization. I asked Codex 5.4 x-high to follow the official guidelines for localization and from that perspective the implementation was good. But then it decides to re-write the copy and layout of all pages. They were placeholders, but still?

Codex also has a tendency to apply unwanted styles everywhere.

I see similar tendencies in backend and data work, but I somehow find it easier to control there.

I'm pretty much all in on AI coding, but I still don't know how to give these things large units of work, and I still feel like I have to read everything but throwaway code.

No it is not, but they had a unique positioning around open-source and the parent commenter means that they are losing it.

Specialization works by vertical or by function. Or you can actually mix them if your TAM remains large enough.

Yes, the parent was referring to technical specialization. But my point is either works. Especially in the context of what OP is trying to do which is "automation" - technically very broad.

I've tried reading this and I can't. It's not that the text is AI generated, it's that the whole structure seems to be. (Hope you appreciate the irony of my LLMisms). It's not human-parseable, at least not by this human. And it's not that my attention is shot, luckily I'm still able to read copious amounts of long-form text and analysis.

Also, opening with "I'm a top performer"... That's not how writing for other humans works. It's perfectly legitimate to establish authority in the opening a piece, but you have to show some credible proof. "I'm a top performer" is immediately off-putting.

I've drifted across projects in different industries (FMCG, investment funds, ad agencies, startups of various sorts) and like I said I had a long project (over two years) for an executive search firm and got to see the ins and outs of how everything works from strategy to technology. I could be drifting to find clients in yet another vertical but I've decided to stay put for at least a few years. So to answer your question, in my particular case: I drifted, stumbled upon something by chance, and then took a conscious decision to stay.

> Become an expert in 1 thing

I endorse this. I've been doing generalist consulting for about six years, and I love flying solo. I've been successful in landing some big customers and interesting projects, but I'm tired of the inefficiency that comes with being a generalist, so I've decided to specialize vertically.

I had a super-interesting project in executive search in the last couple years, and I've decided to settle around that area: executive search and recruitment firms. Maybe later, as an extension I'll target other B2B, relationship-driven professional service firms tha share a common core of processes.

I've only recently pivoted but I'm already starting to see the fruits. It's commercially efficient. Many potential customers seem happy to open the door and chat. I know where to find them, online and off. And then it's operationally efficient. I'm confident I could jump on a customer project and recognize most of their processes and systems immediately and have a quick impact. I already have a base of IP (documented business procedures, code, etc.) and only intend to grow it in the coming years and even turn it into a "productized service".

I think people refuse to specialize for three main reasons. The first is for lack of a clear thesis. That's fine, you need to explore for a bit. The second is for a fear of lack of opportunities, which is often unfounded. The third is due to psychological reasons related to the image of self. On this last one I can only advise that (a) even in specialization there is way more variety than you think, (b) you can always keep growing as a generalist with side projects and self-directed learning and (c) nothing is ever fixed in stone, everything is in flow - you can always pivot out into other interesting directions.

I don't dislike AI, I use it every day for coding and increasingly for non-technical tasks, and have also used it in enterprise workloads to great success. I am fairly optimistic about it - I think it will remove a lot of drudgery and make things economical which previously weren't.

I am just challenging the notions that "if you limit it to text, it's doing really well" or that the text contains in itself all the information that is needed to carry out a task to a certain level of quality. This applies in my experience not only to writing literature but also to certain human tasks which may appear mundane and easy to automate.

I agree that the argument starts from a reduction to the absurd.

So at least we can agree that AI hasn't mastered the text medium, without further qualification?

And what about my argument, further qualified, which is that I don't think it could even write as well as a good professional writer - not necessarily a generational one?

I don't think it can write something that's of a fraction of the quality of Kafka.

But GP's argument ("limit the space to text") could be taken to imply - and it seems to be a common implication these days - that LLMs have mastered the text medium, or that they will very soon.

> it can't write a foundational piece of literature

Why not, if this a pure textual medium, the corpus includes all the great stories ever written, and possibly many writing workshops and great literature courses?

I've glossed over a few and one can immediately tell they don't meet the average writing level you'd see in a local workshop for writers, and much less that of Mann or Capote.

> What's your bar that needs to be passed

I provide four examples in my comment...

> you just limit the space to text

And even then... why can't they write a novel? Or lowering the bar, let's say a novella like Death in Venice, Candide, The Metamorphosis, Breakfast at Tiffany's...?

Every book's in the training corpus...

Is it just a matter of someone not having spent a hundred grand in tokens to do it?

No - that’s my secret.

Oh yea, I agree that building good software remains roughly as challenging as ever.

I was asking if there was something about the “agentic” part in particular that was difficult.

> It’s really hard

How? I just open multiple terminal panes, use git tree, and then basically it’s good old software dev practices. What am I missing?

I’ve read and heard from Semi Analysis and other best-in-class analysts that the amount of software optimizations possible up and down the stack is staggering…

How do you explain that capabilities being equal, the cost per token is going down dramatically?

I am starting to use Codex heavily on non-coding tasks. But I am realizing it works because I work and think like a programmer - everything is a file, every file and directory should have very precise responsibilities, versioning is controlled, etc. I don't know how quick all of this will take to spread to the general population.

I’ve read it, it’s taken me about ten minutes, and it’s been refreshing to read something that is not sloppy. The page design and type are soothing, too.

I don’t agree with everything that the article says but it soulfully blends concepts in history, politics, economics, cryptography and AI.

I don’t think the author could’ve compressed it without precisely sacrificing the essay’s soul.

Is this what everything is coming down to?

How do people like that ascend to CTO?

I would say historical materialism is way more influential. His theory of value was quickly dispelled (although many continued to believe in it) but all marxist and post-marxist thinkers (the Frankfurt school, French post-modernists, current woke academicians) continue to use historical materialism in one way or another.