Link tags: programming

Generative AI Is Not Going To Build Your Engineering Team For You - Stack Overflow

People act like writing code is the hard part of software. It is not. It never has been, it never will be. Writing code is the easiest part of software engineering, and it’s getting easier by the day. The hard parts are what you do with that code—operating it, understanding it, extending it, and governing it over its entire lifecycle.

The present wave of generative AI tools has done a lot to help us generate lots of code, very fast. The easy parts are becoming even easier, at a truly remarkable pace. But it has not done a thing to aid in the work of managing, understanding, or operating that code. If anything, it has only made the hard jobs harder.

Home-Cooked Software and Barefoot Developers

A very thought-provoking presentation from Maggie on how software development might be democratised.

The deskilling of web dev is harming the product but, more importantly, it’s damaging our health – this is why burnout happens – Baldur Bjarnason

A depressing but accurate description of the economics of web development.

The quiet, pervasive devaluation of frontend - Josh Collinsworth blog

It’s like CSS exists in some bizarre quantum state; somehow both too complex to use, yet too simple to take seriously, all at once.

In many ways, CSS has greater impact than any other language on a user’s experience, which often directly influences success. Why, then, is its role so belittled?

Writing CSS seems to be regarded much like taking notes in a meeting, complete with the implicit sexism and devaluation of the note taker’s importance in the room.

I worry our Copilot is leaving some passengers behind - Josh Collinsworth blog

Products of all kinds are required to ensure misuse is discouraged, at a minimum, if not difficult or impossible. I don’t see why LLMs should be any different.

LLMs and Programming in the first days of 2024

What strikes me about my personal experience with LLMs is that I have learned precisely when to use them and when their use would only slow me down. I have also learned that LLMs are a bit like Wikipedia and all the video courses scattered on YouTube: they help those with the will, ability, and discipline, but they are of marginal benefit to those who have fallen behind. I fear that at least initially, they will only benefit those who already have an advantage.

Eigensolutions: composability as the antidote to overfit • Lea Verou

I love, love, love the deep thinking that Lea has put into this, really digging into the guts of what design does.

Overfitting happens when solutions don’t generalize sufficiently and is a hallmark of poor design. Eigensolutions are the opposite: solutions that generalize so much they expose links between seemingly unrelated use cases. Designing eigensolutions takes a mindset shift from linear design to composability.

Lea ties this into web standards too. It’s really helped clarify for me why I want more declarative options for common use cases (like a share button)—it’s about raising the ceiling without raising the floor.

Pipe Dreams: The life and times of Yahoo Pipes

Yahoo PIpes was ahead of its time. Here’s a nostalgic retrospective by Glenn Fleishman.

The subjective experience of coding in different programming languages (Interconnected)

I love the analogies Matt uses to describe the vibes of different kinds of coding:

When I’m deep in multiple nested parentheses in a C-like language, even Python, I feel precarious, like I’m walking a high wire or balancing things in my hands and picking my way down steep stairs.

I haven’t done much Haskell but what I did felt like crawling underground through caves and tunnels.

Opening a terminal window to a distant server is like reaching through a hatch with my arm, but a long way; ssh tunnel is well named.

Writing code with GitHub Copilot and Typescript in full flight feels like, well, flying, or at least great bounding leaps like being on the Moon.

Losing the imitation game

The hard part of programming is building and maintaining a useful mental model of a complex system. The easy part is writing code. They’re positioning this tool as a universal solution, but it’s only capable of doing the easy part. And even then, it’s not able to do that part reliably. Human engineers will still have to evaluate and review the code that an AI writes. But they’ll now have to do it without the benefit of having anyone who understands it. No one can explain it. No one can explain what they were thinking when they wrote it. No one can explain what they expect it to do. Every choice made in writing software is a choice not to do things in a different way. And there will be no one who can explain why they made this choice, and not those others. In part because it wasn’t even a decision that was made. It was a probability that was realized.

This post also has a really good explanation of how large language models work.

There may be real, productive uses for these kinds of tools. There may be ways to build and deploy them ethically and sustainably. But that’s not the situation with the instances we have. AI, as it’s been built today, is a tool to sell out our collective futures in order to enrich already wealthy people. They like to frame it as being akin to nuclear science. But we should really see it as being more like fossil fuels.

A Coder Considers the Waning Days of the Craft | The New Yorker

GPT-4 is impressive, but a layperson can’t wield it the way a programmer can. I still feel secure in my profession. In fact, I feel somewhat more secure than before. As software gets easier to make, it’ll proliferate; programmers will be tasked with its design, its configuration, and its maintenance. And though I’ve always found the fiddly parts of programming the most calming, and the most essential, I’m not especially good at them. I’ve failed many classic coding interview tests of the kind you find at Big Tech companies. The thing I’m relatively good at is knowing what’s worth building, what users like, how to communicate both technically and humanely. A friend of mine has called this A.I. moment “the revenge of the so-so programmer.” As coding per se begins to matter less, maybe softer skills will shine.

The Illusion Of Developer “Productivity” Opens The Door To Snake Oil – Codemanship’s Blog

There are a lot of astute observations in here.

When I lost my job, I learned to code. Now AI doom mongers are trying to scare me all over again | Tristan Cross | The Guardian

Ingesting every piece of art ever into a machine which lovelessly boils them down to some approximated median result isn’t artistic expression. It may be a neat parlour trick, a fun novelty, but an AI is only able to produce semi-convincing knock-offs of our creations precisely because real, actual people once had the thought, skill and will to create them.

AI isn’t the app, it’s the UI - Stack Overflow Blog

In some ways, the fervor around AI is reminiscent of blockchain hype, which has steadily cooled since its 2021 peak. In almost all cases, blockchain technology serves no purpose but to make software slower, more difficult to fix, and a bigger target for scammers. AI isn’t nearly as frivolous—it has several novel use cases—but many are rightly wary of the resemblance. And there are concerns to be had; AI bears the deceptive appearance of a free lunch and, predictably, has non-obvious downsides that some founders and VCs will insist on learning the hard way.

This is a good level-headed overview of how generative language model tools work.

If something can be reduced to patterns, however elaborate they may be, AI can probably mimic it. That’s what AI does. That’s the whole story.

There’s very practical advice on deciding where and when these tools make sense:

The sweet spot for AI is a context where its choices are limited, transparent, and safe. We should be giving it an API, not an output box.

Rich Harris: Hot takes on the web 🌶️ - YouTube

I don’t agree with all of these takes-of-varying-spiciness, but Rich Harris is always worth paying attention to.

Why ChatGPT Won’t Replace Coders Just Yet

I’ve been using Copilot for over a year now, and this is more or less how I use it: To help me quickly blast through boilerplate code so I can more quickly get to the tricky bits.

There’s a more subtle problem with ChatGPT’s code generation, which is that it suffers from ChatGPT’s general “bullshit” problem.

Fragile Technologists – Terence Eden’s Blog

If you’ve made a computer do something - anything - in a logical fashion, you’re a programmer. Scratch? Programmer! CSS? Programmer? Conditional formatting in Excel? Programmer!

Theory-building and why employee churn is lethal to software companies – Baldur Bjarnason

This extract from Baldur’s new book is particularly timely in light of the twipocalypse.

CSS Timeline

Here’s a remarkably in-depth timeline of the web’s finest programming language, from before it existed to today’s thriving ecosystem. And the timeline is repsonsive too—lovely!

Towards Growing Peaches Online - by Claire L. Evans

A beautiful meditation on Christopher Alexander by Claire L. Evans.