d0mine

These are orthogonal:

- there are people who do not want LLM contributions being accepted into emacs itself (legal, maintainers burnout, etc reasons)

- there are projects (such as gptel, pi-coding-agent) that enable LLM usage from within emacs

Both may be true at the same time.

How many accounts Anthropic used to mine the web

It is more like an ability to recognize the word when it is used in context.

I got ~1/3 that is very generous estimate even for "recall" case (recognize), and it obviously false for the "generate" case (using in speech) where I guess my vocabulary is likely ~1/90 of all English words.

Mimocode(opencode), Pi Coding Agent

It is how these agencies operate. They mask malevolent activity behind good front (“think of the children“, “age verification“ -> censorship/total surveillance. Medical aid -> overthrowing governments).

As I understand RL makes foundation models stupider (less capable, not more) but better at following instructions.

U.S. Military Turned GPS into a Global "Numbers Station" https://news.ycombinator.com/item?id=48414479

In practice, when conventional commits are used with git emojis, they look like “scope: what is done” already (“<emoji> <issue-id> scope: …”)

Conventional commits (especially with git emojis) show at a glance the blast radius of the change (eg whether it breaks the product itself or just some internal dev tools). Emojis help immensely when looking at dozens of commits at a time.

If you need a network boundary, what MCP provides that REST API + llms.txt can't do?

In practice, you would use an already written implementation, maintained by somebody else. An option that is often ignored by LLM (copy-paste galore).

For example, imagine if textual-serve author would reimplement xterm.js What effect it would make on quality.

LLMs increase technical debt rapidly. It is unclear whether they can deal with the mess they create. But we'll know soon (no need to wait years, to get immovable mess).

The positive side of LLMs is that they confirm experimentally the usefulness of many software engineering practices (testing,docs, adrs, design, formal specs etc)

Almost everything you see on the topic is propaganda. Alternative views are suppressed (though I’m surprised on the relative diversity I’m seeing in this thread)

“Either or” is too extreme. Both can be true at the same time even if at different levels of abstraction eg a table and atoms it consists of may exist at the same time. Depending on the task different levels of abstraction may be useful. It is ridiculous to claim that a table doesn’t exist just because we can prove it is made of atoms (or strings excitation—doesn’t matter here)

On the other hand, "consciousness" concept might be as much useful for modeling thinking as “the four elements” for describing anatomy (not useful at all)—and we create better models eventually.

“Those who say they lack time to build tools are precisely the ones who cannot afford not to.”

Unlike Kindle, Amazon can’t shutdown your bookshelf

“Parse don’t validate“ seems like the same idea

https://lexi-lambda.github.io/blog/2019/11/05/parse-don-t-va...

You do not need Haskell for that eg it works in Python (via pydantic, attrs data classes)

Here's github statistics from 2015 https://github.blog/open-source/open-source-license-usage-on...

MIT is used by more projects than GPL.

What part of the reply is not factually correct? (premium gold+ open access article for $50000?)

You are called terrorist only until you win then you are a freedom fighter.

You even may be called freedom fighter from the start if you are trying to displace government in the right country. There are plenty of examples.

how our brain construct

is the only part i.e., we perceive what brain predicts no more no less. Optical illusions demonstrate it well.

Sometimes that prediction (our perception) correlates with the light reaching the retina. But it is a mistake to think that we can perceive it directly. For example, we do not see the black hole in our field of vision where there are no receptors (due to our eyes construction).

Another example that makes the point clearer: there are no "wetness" receptors at all but we perceive wetness just fine.

You are describing gerrit.

https://www.gerritcodereview.com/

Trying to compress the code may be a proxy for the complexity metric

To get perspective(we know what worked), here’s some 50+ years abstractions:

A file is a simple stream of bytes in Unix. (If you think what else it might be then compare to Multics’ segments). Separate processes that may be connected using simple standard I/O streams [pipe] (vs everything is DLL in Multics) — the concept of shell itself (policy vs. mechanism separation http://www.catb.org/esr/writings/taoup/html/ch01s06.html ).

https://retrocomputing.stackexchange.com/questions/15685/wha...

For comparison, you need a new app on iOS for what might have been a shell pipeline (hierarchical file system is absent at user level).

The downfall was long before ChatGPT (personally, I’d burnt out answering the same questions again and again) https://data.stackexchange.com/stackoverflow/query/1926661#g...

The stackoverflow was better than what was before it (*sexchange site). Time will tell how it compares to the AI slop of the dead internet (current generation of LLMs based on human-generated data are great. Let’s see what happens when most of the data is created by bots).

I don't get how camp 1 can ship more than one version (do they jump teams/companies each time?). If your code is immovable mess then how do you add features/fix bugs in time?

There is a considerable overlap between people who believe that a reality tv personality has real power and those who believe that reality tv is unscripted.

related: "ChatGPT and AlphaFold Help Design Personalized Vaccine for Dog with Cancer" https://www.the-scientist.com/chatgpt-and-alphafold-help-des...

16000 layoffs cost just $180 million—it is a win so far

https://x.com/gothburz/status/2031778265958842541

There are browser extensions such as Vimium C that provide keyboard-based navigation.

`ollama launch opencode --model glm-4.7-flash`

https://ollama.com/library/glm-4.7-flash