submeta

Will these models eventually replace all knowledge work, leaving lawyers, doctors, product managers, software developers, and others out of a job?

If the benefits were shared across humanity, that could bring us closer to utopia. My worry is that we’ll instead end up with a handful of even wealthier billionaires and millions of people out of work.

I went from IDE (VS Code) to terminal (wezterm) running half a dozen agents in an app called Herdr. My computer work is mostly in the terminal now, prompting agents. And they have access to my Obsidian vaults. Can read, and add to them. And access all my notes via a local RAG. I neither need projects in Claude, nor IDEs. Just agents accessing my local vaults and code repos.

I don’t understand the downvotes. It’s perfectly reasonable to disagree, but the argument is laid out clearly. Why not engage with it instead?

Some 20 years ago, the telecommunications sector in Germany was liberalized. Many telephone card providers entered what had previously been a barely competitive market. They advertised their products with aggressive claims like: “Buy our €10 top-up card and get 660 minutes to destination X.”

For the first few weeks, they would actually provide those 660 minutes to establish trust in their cards. But after a while, they would quietly start reducing the number of minutes on subsequent top-ups—say, from 660 minutes down to only 300. They wouldn’t do this for every card, so it was difficult to prove. Instead, they relied on averages across their customer base to make the economics work.

Lately, I’ve found myself wondering whether something similar may be happening with frontier AI models. Companies launch with an exceptionally strong model and generous compute limits to build adoption. Once the model is established as a market leader, the incentives change, and users may start perceiving the service as becoming more constrained or less capable over time.

I don’t have evidence that this is what’s happening with Anthropic—or with any other AI company. It’s simply a pattern that the current situation reminds me of.

Which prompt, though? Just the initial one, or the entire conversation? During a project you’ll prompt the agent dozens of times. Are all of those prompts considered the source?

For those who moved away from GitHub: do you miss the social/discovery side of it?

I mean things like comments on issues/PRs, stars, followers, finding existing repos, seeing which projects are popular, and getting drive-by contributions.

Or does that matter less than I imagine once you self-host and mirror public repos back to GitHub/Codeberg?

I‘d go with Scheme. You‘ll learn the basics in a day. The language spec is only a few pages. And Scheme reads like pseudo-code with parentheses.

Wouldn’t a modern app like iA Wtiter or any other distraction free full screen text based writer be just as good? Or why does WordStar still have his fanbase decades later?

What is the incentive of the U.S. government? Is it to prevent adversaries from accessing this powerful technology? Because of security concerns? Or is there a fear that European, Asian, Latin American, and other companies could use it to build competitive products? What happened to free trade? What about all of humanity advancing and making progress?

Are GPT 5.5 and Opus 4.8 the last models we're going te be allowed to use in Europe? Is there going to be a cut, and we're only be allowed to use less capabale models outside of the US?

I mean, if they deem Fable 5 to powerful to share with the rest of the world, what's left for us?

What?

M5 Max MacBook Pro: $4,099 (up from $3,599)

M3 Ultra Mac Studio: $5,299 (up from $3,999

How can this be explained with price increases in Ram prices?

Come on Apple, don’t be so greedy. Make money but don’t bleed us.

I suspect AI is replacing my need for productivity content much faster than it’s replacing my need for books.

I read fewer blog posts, fewer newsletters, fewer “10 lessons from…” articles, and fewer productivity videos than I did three years ago.

But I still buy books.

The first casualties seem to be the intermediaries, not necessarily the original sources.

Congrats to the Cursor team. Unbelievable success. Can’t imagine how their live suddenly will change.

First they came for the Chinese (can’t buy GPUs), then all of non US citizens (can’t use latest models). What’s next, we can’t use encryption? Cybersecurity tools? Access to latest publications in science?

Can this be used as an alternative to Claude backend? For Ralph loops? Replacing `claude -p`? Anyone can shed a light on this?

Love it! Thanks for sharing. I use mermaid a lot, and this I will use to embed my diagrams as ascii in my text docu.

Absolutely love HP calculators. Almost fourty years ago, my HP28s introduced me into programming in RPL, a stack based Forth like language combined with elements from Scheme (Lisp), and symbolic calculation. It was a language that was way ahead of its time, nothing I had seen in Pascal or C. Only a few years later in Mathematica. From there I learned Scheme, and was introduced to the book SICP. All of this had a lasting effect on how I program and think.

HP started my journey so to speak :)

I also had one of those mentioned in the article, just for nostalgia. Rock solid, RPN based, lovely product. The kind of product companies do not build anymore (products that will last you a lifetime)

I cannot imagine that students, kids, or half the world where paying 3 usd a month is impossible, will keep using whatsapp when they have to pay a fee. They will look for alternatives immediately. Telegram?

But actually this is a good move. I tried to convince my family and friends to use alternatives, without success. But now I see hope.

Have stopped using FB and IG years ago, was stuck with WhatsApp because of half the world using it.

I used a predecessor of this almost thirty years ago to learn Scheme and work trough the book SICP. The Racket maintainers still ship updates and new features, that’s remarkable.

Scheme is a wonderful lisp dialect. It taught me basics of functional programming, about closures, about tail call recursion, about functions always returning values (which annoyed me a lot when I started learning Python, where .append or .sort returened `none` instead of the list, and were destructive).

So I have very fond memories of Racket (then DrScheme) and Scheme. Had also written my matrix multiplication library and my CAS system to mimic the functionality of my HP28s calculator.

Have to look into it again.

Companies might need Microsoft, but why are people panicking who could replace ms office with other office suites? Why aren’t they abandoning Microsoft products? From office suites to windows?

Not lost because it's hard to learn, but because I don't like writing in ms office products. It's not just word, I write formated long emails in outlook as well.

I am a heavy user of Pandoc. As I write all my text in markdown using Obsidian, but have to create content for the MS Office environment, I use Pandoc to convert my markdown content into ms office formated content.

I would be lost had I have to use the Office tools to edit and format my text.

So thank you to all the maintainers of Pandoc.

The author is blaming llms alone. And is not even mentioning sites where you can download any and every book you wish. That might have led to decreasing sales as well.

Not everything can be answered with yes or no. I used to give thorough answers even pre-LLMs. When someone asks about a project and its details, I am supposed to give a one sentence answer because LLMs give lengthy answers and that sucks?

This is like people hating on em-dash because LLMs use them a lot.

BBEdit used to be my text-transformation tool.

Happily paid for every update for years, even when I used Emacs, I kept BBedit in reach. For quick text edits/transformations (because Regex in Emacs is hard to use). But with LLMs + nvim I hardly start bbedit anymore.

So now with LLMs, I tell them what I need and they write a shell/Perl/Python script to make the craziest transformations.

100%. This is what’s happening with agentic coding. We develop individual solutions tailored to our needs. We don’t need one-stop shops anymore.

For instance I have created a snippet manager, because I didn’t like the one in Raycast. Or I have created a journal viewer because the daily notes in Obsidian are not easily viewed across days. And so I have dozens of smaller solutions that are tailored to my needs. And I am guessing lots of vibe coders do the same.

This is true. At our company they rolled out ChatGPT with Codex. After two months of happily using it, I got a call from the IT OPs telling me I burnt through four hundred million tokens, 200m a month. And created at least a thousand euro bill. That’s after I used all the credit, but I don’t have all details. The guy told me to „watch my usage.“ What does that even mean. He doesn’t use it himself and apparently he doesn’t know how value is created here and how he can monitor and limit usage.

Did OpenAI switch from fixed prices per seat to usage based? This will surprise many companies I reckon.

Personally I use Claude Code, the 200 euro plan. And am a heavy user. A few weeks ago I realized that CC shows the token usage in cli, in the bottom right. Something I never cared about because I thought paying 200 euro a month will give me „unlimited“ access.

But I guess the party is slowly coming to an end? Prices are going to increase slowly? And the flatrates will be removed eventually?

Too bad, it was nice while it lasted.

And he doesn’t use any config it seems. No syntax highlighting, no line numbers, plain white text on dark background.

I use Claude Code, the max plan, for 200$ a month. Privately. So if I convince my peers at work to use this, will it cost significantly more? Because it's token usage based? Not flat rate? And if so, how much more? I think I burned through 400m tokens in two months in Codex. Can someone shed on light on this?

There‘s HP calculator guys and TI guys. Around the age of 17 I spent lots of time programming my HP28s calculator in a Forth like language that had symbolic mathematics, lots of ideas from Scheme (closures, functions as first class arguments, recursion). It felt like magic dealing with concepts I hadn’t seen in the C compiler on my Amiga or later in Turbo Pascal. But I saw these concepts later in Mathematica and was familiar.

I had programmed games, complex 3d visualisations (super slow but oh well), and was totally fascinated by what this device could do.