bandrami

Dunbar strikes again

I think the argument is the Indus Valley civilization was dependent on international trade with the autocracies further west and collapsed when that network broke down

I'm sorry but it's hard not to see this as a tool mainly designed to inflate token counts

There's an alternate timeline where Ousterhout's dream was realized and TCL filled the role that the Javascript swamp filled. I think that's a better timeline from a tech perspective.

Is it the models or the users' dopamine receptors that get nerfed after a few weeks?

Ken Lay would like a word

The fact that Ellison just backed out of a massive stock sale makes me think everybody is reading the same bond market numbers

That's "rude" but allowed; that capability is one of the reasons Wayland was developed (though their solution removed some capabilities people wanted with the result that the compositors now replicate the original flaw).

Everybody's out of money and they need a way to calmly unwind this. Same reason Ellison just backed out of the stock sale.

A chimpanzee could turn the nuclear keys if he were given access to them. The solution is not to remove chimpanzees' thumbs; it's just to not give them the nuclear keys. This is a very basic lesson that the labs are somehow missing.

Every one of those posts costs money, and OpenAI is apparently willing to pay that. This really is a problem with a simple solution.

iPython was famous for a while but I think its popularity has waned, but yes as you point out really any dynamic language can do a repl; it's just that it's mostly the lisp/scheme and forth families that have leaned so heavily into the idea.

It's the same reason people like symbolic debuggers: you can run your program as many lines at a time as you want and stop and inspect the state whenever you need to. I haven't really used jupyter but I understand the workflow there is very similar.

As a guy who started in lisp I always find LSP to be a weird uncanny valley situation. It's like 99% of a REPL, but it will always have corner cases where that 1% difference is important.

Commodification of inference means the death of capital spending on buildouts, which would mean the end of both Softbank and probably Oracle.

Aren't the "thinking" chains always just reconstructed anyways? It would be like using a debugger that just looks at the source code rather than the actual binary.

There are whole industries that still don't let LLM-generated code touch production

I gave up on gun control as a political question when I realized you can buy a flamethrower at Home Depot

To what extent is OpenClaw still alive to be killed?

Or, I mean, reading and responding to emails on your phone directly

The only job application process I've ever walked out on in the middle of was 25 years ago, when I was taking a proctored computer skills test and it dinged me for pressing the Windows key rather than moving the cursor to the Start button and clicking it.

So this is a new presentation of an old problem, I guess is the takeaway there.

IDK, it's the same problem I had with OpenClaw: I can't think of anything I need done that I want it to do for me.

A year ago everybody in San Francisco was saying the entire world was about to be blindsided by the fundamental changes that would happen in the next 8-12 months and, well, that just hasn't happened.

Hard disagree, as someone whose main job is telling developers "no". If it's everybody's responsibility it is in fact nobody's responsibility.

It has eval but it lacks reader macros (or at least it did thirty years ago) which is what makes eval useful on lisp2 systems.

Somebody's got to actually operate those robots, after all

Particularly since one of their flagship products is a CMS that solves that problem and lets you move documents around without breaking links

Meta (née Facebook) hired addiction medicine specialists to work on their platform engagement. It wasn't to make teens use their platforms less.

Smart speakers are a mature market and seem to have topped out at about 1/3rd of US households at a lower price point than this. This is going to be a very hard sell.

I still think this question is a future minefield, as well as the question of what happens if an LLM outputs a patented algorithm