warrenm

Alternative link: https://archive.is/ZzD9m

Most of the self-proclaimed 'emperors' have no clothes

The actual emperors? They are dressed (or, at least, in the fabric section of the hobby store and have a sewing machine at home)

>why did they try so hard to hype it in anthropomorphic terms?

Marketing, pure and simple

People relate better to something that sounds humanesque (even though it is not) vs calling it what it is (in this case, a massively-backed (ie LLM-based) Markov Chain generator)

>How close are we to where a robot could get into a car from 2010 and drive me around?

A long way away

And here is why - driverless cars are a thing ... essentially making the car the "robot"

General-purpose robots are an amusing scifi trope, but have no practical benefit in reality

Purpose-built robots (even ones that can flex within that prupose to different applications) make far more sense (and have already been around for decades)

This is precisely why nVidia is providing GPU to AI companies, and not trying to be one themselves ... loads more money to be made selling shovels to prospectors than in being a prospector

The valid use cases for blockchain are relatively few

The use cases where it gets applied are far more than the valid ones

As for NFTs ... there never was (and never will be) a valid use case - it does not matter if you "own" a digital asset (like an image): a screenshot of it is good enough for 99.99999...% of people, so why pay for the "real" thing?

Define "malicious code"

Now define "unintended side effect"

Now add "no one is maintaining it anymore"[0]

-------

[0] https://xkcd.com/2347/

And that

how much of a propulsion system could you feasibly pack in a cubesat?

Sounds like you have yourself a YCombinator startup proposal in the making

>Yep, but students love reinventing the wheel ;).

And ... professors love making students reinvent the wheel

GPUs are [effectively] irrelevant for many use cases (IoT, embedded, most servers, etc)

I was on the Fediverse for a while with my own Mastodon instance a few years ago[0]

Then a minor update broke it[1]

Have not been back since

Like you, I want to like it / use it ... but it simply takes too much mental space on top of the other social media / comm channels I use (professional and personal)

--------

[0] https://antipaucity.com/2019/02/07/sweetree-ga-the-newest-ma...

[1] https://antipaucity.com/2019/07/16/goodbye-self-hosted-masto...

It is 'worth' whatever someone will pay for it

No more, no less

How much work you put into it matters not to a buyer

All that matters is how much value a buyer thinks it will yield them

So ... they are moving to a more-accurate pricing model, and you are upset?

Some things cost less than others

AWS (and every other cloud provider I can think of) does this with their pricing - bandwidth (fully-internal, cross region, external...), compute (EC2, Lambda...), storage (S3, EBS, local NVMe...), and more

Sounds like you have now gotten a clearer mechanism to understanding usage, so you can maximize your utilization, and you want to have an opaque view instead

If I were to get a "coding assignment", I would tell them 'no'

So I would spend 0 minutes

And not just because I am not especially close to a 'coding' or 'developer' role in my IT career

But because they are all* a complete and utter waste of time

Unless you mean a whiteboarding session in an in-person interview, where you see approaches to problem solving, swags in the direction of an answer, refinement of the problem statement, etc - those I spend 15-20m on (out of the 45-60m interview session)

---------

* maybe there is a coding assignment that is not a waste of time, but I have never in my well over 20 years in IT seen it

Airlines make their money - primarily - on cargo and fees

Not on passenger tickets

Starting in the middle, vs one end or the other, is definitely a different problem :)

you do not need to remember state with the simplest solver:

- place your right hand on the right wall - walk forward, never letting your hand leave the wall - arrive at the exit

yes, you travel many dead ends along the way

but you are guaranteed to get to the end of a 'traditional' maze

>I've tried having them run a maze, but instead of giving them the whole maze up front, I have them move one step at a time, tell them which directions are open from that square and ask for the next move, etc.

Presuming these are 'typical' mazes (like you find in a garden or local corn field in late fall), why not have the bot run the known-correct solving algorithm (or its mirror)?

You can "self-host" Signal!?

Jitsi supports it - https://jitsi.org/e2ee-in-jitsi/

Full report: https://washdata.org/reports/jmp-2025-wash-households

Tragic stories like these underscore how absolutely vital it is that parents be involved in their kids' lives - offline and online!

I can not begin to imagine how devastated his family is over all of this - how many ways and times they will ask themselves, "what could we have done differently?"

LLM-based tools like ChatGPT and Grok etc can be fantastic - when used well

And absolutely horrible if used badly

As they statistically respond to prompts and responses, spitting out text from their vast libraries, negative prompting will - inevitably - lead to negative text generation

You have probably heard of the 7-38-55 Rule [0] (some round and call it the 60-35-5 rule) which observes well under 10% of communication is via the words used - body language, the biggest component, is completely absent in a tool like a chatbot. Tone is [nearly] undeterminable in pure text - take, for example, Harrison Ford reading the opening words to "The Cat in the Hat" in Patriot Games while his daughter is in the hospital after the car crash: that is a "fun" book, yet the way the opening is read is very ominous.

Chatbots, like text messages, IMs, email, etc, have tone assumed by the other party: if the person interacting is in a good mood, even if talking about sad/negative things, the inferred tone in the bot's responses are going to be very different than if the person interacting is in a foul mood of some kind.

Combine that with how relatively easy it is to manipulate the prompts to get different results (even when "asking the same thing" (eg "how many b's are in blueberry" vs "how many letters b are in the word blueberry" [1])), and these generative engines - while, again, amazing tools - are triggered through their algorithmic matrices to produce whatever they have been 'trained' to produce.

-------

[0] https://www.masterclass.com/articles/how-to-use-the-7-38-55-...

[1] https://news.ycombinator.com/item?id=44888359

via https://flowingdata.com/2025/08/25/there-is-no-correct-map/

Sounds like how pretty much all politicians 'worry' about things that cannot occur until years after they are out of office

Was Clancy an "insider"?

No

He just understood how to ask for non-classified information, collate various and sundry ideas, and write well

>It’s misleading to speak of a wallet as holding cryptocurrency. It holds private keys that allow you to access cryptocurrency, so in that sense it’s a password manager. Not only that, these days it’s effectively a password manager containing only one password.

"Experts" in any field hate being told when they are wrong, or how they could be better - it upends their sense of self-importance of being the "expert"