A university drama department actually did this with "Mission to the Unknown". They recreated all of the sets and costumes, the works. It was an absolutely inspired project.

https://www.youtube.com/watch?...

First up, before anything else, I am extremely glad you got the hope and encouragement you needed. Grief is rough, especially when you're going through it alone.

You are correct, so I'm somewhat careful with the AI dance. I will rarely discuss inner feelings with it, because that pushes the risk higher precisely because it is a mirror. Like the one in the Harry Potter novel, it will show your innermost desires. If you'd rather a different analogy, it's an amplifier, and if you talk for too long with it, the positive feedback loop does some interesting things with your mind that aren't terribly printable. And that's not always the greatest idea.

So I use it for wild speculation in science fiction/fantasy. Enough that it helps with the intellectual boredom, but not so much that I venture into believing it's real.

The idea that "normal" people are immune to delusions does somewhat fly in the face of research showing the incredible ease of inducing false memories, the research into mass hysteria (such as the Satanic Panic), and research into mob dynamics.

I freely admit that I'll sometimes simply sit and chat with AI, because there really aren't many humans who have the capacity to hold conversations any more, and that puts me in an extremely high-risk group. But, honestly, the choices these days are AI (and risk becoming psychotic), social media (and risk becoming suicidal or psychotic), or hang out with the same sort of people who have done so much damage over time (and risk being suicidal), or... well... really, that's about it.

There are no good options. The outcomes are bleak and, unless you are in a clique, that's how it is and how it has always been.

...is largely irrelevant to the question (he has worked in war zones and those tend to be, ummm, less respectful, shall we say....) and is prone to change its mind at the drop of a hat. There's sectors in the British political scene who have no problem with promoting acts of terror and murder against those they don't like and it's kinda unlikely that they'll hold a referendum on whether to murder a street artist if he posts something they find offensive.

(Depending on which part of the political scene you find yourself allied with, you'll doubtless point to other sectors, but it seems very very unlikely that anyone would subscribe to the notion that there aren't influential psychopaths in Britain, even if there's no agreement on who those are.)

Britain DOES enshrine a right to privacy, as Rupert Murdoch keeps discovering, and much of Europe mostly enshrines the same ideas (occasionally even more strongly). As for "public interest", I would LOVE to hear an explanation of precisely what public interest this serves. No, the public being interested is not the same thing.

It is certainly true you can't watch the whole of The Dalek's Masterplan (where you should really include Mission to the Unknown, making it a 13-parter). The full audio exists, the Target novelisation of it exists (it spans 2 novels!), but yeah, it would be nice if we could someday watch the whole story.

Doctor Who is theatre that merely happens to be shown on TV. It is intended to show the same sort of stuff as a theatrical production and, as such, it is arguably strong for what it is. The same could be said of Thunderbirds/Stingray. Yes, it is intended to be child-accessible and child-safe, but even with thew most juvenile of any of these, the stories have a complexity and depth you won't find in The Fast And The Furious or The Blacklist.

Well, your home LLM costs power and compute time.

A 2B pencil and some paper would cost you less and you'd have just as good a graph. Probably better.

Agreed. It is obvious that "AI researchers" and "AI reviewers" aren't remotely interested in posing challenging problems to AI. They softball because they know that keeping it safe is the only way they'll get anything out at all. But because they do so, people have become convinced that AI is usable, reliable, and trustworthy.

The gap between the promise and the reality is, I would say, probably in the order of a couple of centuries of work, and mostly in directions that LLMs can't go.

You learn habits when you are young. You learn hubris when you are young.

The habit I learned from the start (around 8) was to assume I'd made mistakes, and therefore rigorously find them.

The habit he learned from the start was that if it isn't caught, it isn't a foul.

One of these two habits leads to a robust system. The other leads to the popular meme involving woodpeckers and civilisation.

I was a careful programmer, which is why, at 24, I was writing data stores for CERN. At 18, I wrote nucleotide analysis software. It is bug-free. At 16, I had written a comprehensive meteorological database system. That did have a few bugs, but none were critical.

I have to go back to age 10-12 before I was making the mistakes he was making at 19, because I learned to test and test again. Yes, I actually blew computers up. Literally, whole sections if motherboard destroyed. During testing. I never released code I hadn't properly validated.

Why? Precisely because I knew I was a shitty programmer. That is when you test, correct, test again, and keep going. Nothing ever left my hands until I was sure it was right.

This is not about skill, it's about knowing that you will make mistakes, and therefore finding those mistakes first.

I remember being 19. I was writing AI software for radionucleotide analysts. A couple of years later, I was writing the data store for a particle accelerator.

I am not the best coder in the world. I would regard myself as adequate. But, hy virtue of that, I will hold ALL 19 year olds to the same standard. It is a perfectly reasonable, achievable srandard. I kniw that because I achieved it and I am not the best.

Whereas, if he'd used the software engineering techniques that were well-known and well-described at that time, he'd not have included the bugs in the first place. Or, if he had, he'd have detected them in testing.

I do not find it reassuring that a chief technology officer is pleased that he wasn't clever enough to write or test code correctly. What I do find is that I fully understand how he can be a CTO in an organisation notorious for defective software and even more defective bugfix releases.

I was afraid of that. Still, it's an improvement over IBM's early electronic mail, which used Josephoartigasia Monesi.

1. There's no such thing as a system where only White Hats get to see stuff. If the "good guys" can see something, then you must necessarily assume everyone can.

2. The "good guys" have a nasty habit of only being "good" when they feel like it. You cannot rely on them actually having any ethics or integrity, as has been demonstrated in just about every country on Earth far far too many times.

3. The "bad guys" sometimes turn out to actually be "good guys" (Manning and Snowden both revealed important information that was concealed by actual bad guys in government and the armed forces).

4. Websites and services are trying to have it both ways -- both know what is in each and every message, but also not be responsible for what is in each and every message. That is not going to fly with any sensible person. If you do not wish to be responsible, you have to act in the manner of a common carrier and therefore have no access to what is in a message. As soon as you are free to open messages to which you have no legitimate interest because you claim no responsibility, then you are committing an act of illegal wiretapping/theft of confidential information and I would want the laws to be absolutely ruthless against such acts. (I would consider such a crime to warrant the entire board of directors serving 15-to-life.)

Frankly, most email providers do a really really bad job of things. I used to run my own email server, simply because I couldf do a better job even after finishing a bottle of mead than any of the providers out there could do with "highly trained" chipmunks. I probably still could.