Opus 4.5 is currently helping me write a novel, comprehensive and highly performant programming language with all of the things I've ever wanted, done in exactly my opinionated way.

This project would have taken me years of specialization and research to do right. Opus's strength has been the ability to both speak broadly and also drill down into low-level implementations.

I can express an intent, and have some discussion back and forth around various possible designs and implementations to achieve my goals, and then I can be preparing for other tasks while Opus works in the background. I ask Opus to loop me in any time there are decisions to be made, and I ask it to clearly explain things to me.

Contrary to losing skills, I feel that I have rapidly gained a lot of knowledge about low-level systems programming. It feels like pair programming with an agentic model has finally become viable.

I will be clear though, it takes the steady hand of an experience and attentive senior developer + product designer to understand how to maintain constraints on the system that allow the codebase to grow in a way that is maintainable on the long-term. This is especially important, because the larger the codebase is, the harder it becomes for agentic models to reason holistically about large-scale changes or how new features should properly integrate into the system.

If left to its own devices, Opus 4.5 will delete things, change specification, shirk responsibilities in lieu of hacky band-aids, etc. You need to know the stack well so that you can assist with debugging and reasoning about code quality and organization. It is not a panacea. But it's ground-breaking. This is going to be my most productive year in my life.

On the flip side though, things are going to change extremely fast once large-scale, profitable infrastructure becomes easily replicable, and spinning up a targeted phishing campaign takes five seconds and a walk around the park. And our workforce will probably start shrinking permanently over the next few years if progress does not hit a wall.

Among other things, I do predict we will see a resurgence of smol web communities now that independent web development is becoming much more accessible again, closer to how it when I first got into it back in the early 2000's.

▲

Madmallard 2 days ago | parent | next [-]

Unfortunately what likely will happen is that you miss tons of edge cases and certain implementations within the confines of your language will be basically impossible or horribly inefficient or ineffective and precisely the reason for it will be because you lack that expertise and relied on an LLM to make it up for you.

▲

soulofmischief 2 days ago | parent [-]

That's not how this works. Assume less about my level of expertise. By the end of a session, I understand the internals of what I'm implementing. What is shortened is the search space and research/prototyping intervals.

If I didn't ultimately understand where I was going, projects like this hit a dead end very quickly, as mentioned in my caveats. These models are not yet ready for large-scale or mission-critical projects.

But I have a set of a constraints and a design document and as long as these things are satisfied, the language will work exactly as intended for my use case.

Not using a frontier model to code today is like having a pretty smart person around you who is pretty good at coding and has a staggering breadth and depth of knowledge, but never consulting them due to some insecurity about your own ability to evaluate the code they produce.

If you have ever been responsible for the work of other engineers, this should already be a developed skill.

▲

Madmallard 2 days ago | parent [-]

Are you making a DSL then? That would make more sense.

	▲	soulofmischief 2 days ago \| parent [-]
		What I am building doesn't work as a DSL, because it relies on compiler optimizations not available to DSLs in other languages. It also has low level support for cross-platform GPU programming. However, I do have support for FFI and also plan to experiment with a WASM port that works with a JS/TS API.

▲

jolt42 2 days ago | parent | prev | next [-]

Long-term maybe we won't care about code because AI will just maintain it itself. Before that day comes, don't you want a coding language that isn't opinionated, but rather able to describe the problem at hand in the most understandable way possible (to a human)?

	▲	soulofmischief 2 days ago \| parent [-]
		You're reading too much into what I mean by "opinionated". I have very specific requirements and constraints that come from knowledge and experience, having worked with dozens of languages. The language in question is general-purpose, highly flexible and strict but not opinionated. However, I am not experienced in every single platform and backend which I support, and the constraints of the language create some very interesting challenges. Coding agents make this achievable in a reasonable time frame. I am enjoying making the language, and I want to get experience with making low-level languages. What is the problem? Do you ever program for fun?

▲

lawlessone 2 days ago | parent | prev | next [-]

Why would anyone buy the novel?

▲

welpo 2 days ago | parent | next [-]

I misread too. "novel" is being used as an adjective, not a noun.

They are saying they are writing "a novel […] programming language", not a novel.

▲

renecito 2 days ago | parent | prev [-]

I'd guess some people likes to read ¯\_(ツ)_/¯

▲

lawlessone 2 days ago | parent [-]

i know, there an inexhaustible amount of human written books to read before i'd be desperate enough to read the Markov chain books.

▲

aoeusnth1 2 days ago | parent [-]

I’d start by reading the comments you are replying to.

▲

lawlessone a day ago | parent [-]

d'oh

▲

soulofmischief a day ago | parent [-]

It happens :)

On that note though, the other day I asked Opus to write a short story for me based on a prompt, and to typeset it and export it to multiple formats.

The short story overall was pretty so-so, but it had a couple of excellently poignant quotes within. I was more impressed that I was reading a decently typeset PDF. The agent was able to complete a complicated request end-to-end. This already has immense value.

Overall, the story was interesting enough that I read until the end. If I had a young child who had shown this to me for a school project, I would be extremely impressed with them.

I don't know how long we have before AI novels become as interesting/meaningful as human-written novels, but the day might be coming where you might not know the difference in a blind test.

▲

lawlessone a day ago | parent [-]

i am in the process of finishing up a role doing annotations for these, for a company i cannot name (basically clicking lots of box hundreds of times a day)

So the endless hosepipe of repetitive , occasionally messed up, requests has probably not helped me endear myself to them.

Anecdotally having chatgpt do some of my CV was ok but i had to go through it and remove some exaggerations. The one thing i think these bots are good at is talking things up..

	▲	soulofmischief a day ago \| parent [-]
		Yes, as it stands now, all frontier models are still downright corny. But a lot of elements of good storytelling are there: the story Opus generated used symmetry and circular storytelling, created tension and release, used metaphor appropriately and effectively... all of those things are there. But the actual execution was just corny. But you should read the stuff I wrote when I was young. Downright terrible on all accounts. I think better training will eventually squeeze out the corniness and in our lifetimes, a language model will produce a piece that is fundamentally on par with a celebrated author. Obviously, this means that patrons must engage in internal and external dialogue about the purpose of consuming art, and whether the purpose is connecting with other humans, or more generally, other forms of intelligence. I think it's great that we're having these conversations with others and ourselves, because ultimately it just leads to more meaningful art. We will see artist movements on both sides of the generative camps produce thought-provoking pieces which tackle the very concept of art itself. In my case, when I see a piece of generative art or literature which impresses me, my internal experience is that I feel I am witnessing something produced by the collective experience of the human race. Language models only exist because of thousands of years of human effort to reach this point and produce the necessary quality and quantity of works required to train these models. I also have been working with generative algorithms since grade school so I have a certain appreciation for the generative process itself, and the mathematical ideas behind modern generative models. This enhances my appreciation of the output. Obviously, I get different feelings when encountering AI slop where in places where I used to encounter people. It's not all good. But it's not all bad, either, and we have to come to terms with the near future.

▲

flanked-evergl 2 days ago | parent | prev [-]

Helping you do something that nobody should be doing is not really compelling.

	▲	soulofmischief 2 days ago \| parent [-]
		Did you have a specific criticism?