To put things in context, this is what they used for communication between a tank and its commanders in WWI:

When the Titanic sunk in 1912, they had a telegraph on board, but no voice radio.

In the 1920s radio took off as a one-way broadcaster to receiver technology, but it still was only rarely used as two-way communications. That only really started for communications between ships in WWII.

So, although she didn’t know how to use the radio in her plane, it was mostly because radio communication was a brand new thing. I’m sure what they put in her plane wasn’t some off-the-shelf radio that had standard switches, antennas and parts. It was probably cobbled together from various parts and only the truly tech-oriented people understood it.

There’s only one little jedi, and it’s on board.

It will slightly improve the chances. But, is that enough?

Imagine you had an intern working with you on a project. They didn’t know anything about SQL injection, cross site scripting, etc. You probably wouldn’t give them a task where that was a concern. If you did, you’d watch them like a hawk. Because they’re an intern, the amount of code they’d produce would probably be pretty low, and it would be pretty low-quality overall, so it would be easy to spot mistakes that would lead to these kinds of vulnerabilities.

An LLM has the understanding of the problem space that an intern does, but produces vast amounts of code extremely quickly. That code is designed to “blend in”, i.e. it’s specifically trained to look like good code, whether it is or not. Because of “vibe coding”, people trust it to do all kinds of things, including implement bits where there’s a danger of XSS or SQL injection. And the way Claude Code ensures it doesn’t generate those vulnerabilities is… someone says “hey, don’t do that, ok?”

Having that statement in there is better than not having it. But, it’s just a reminder that these things aren’t appropriate for writing production code. They don’t actually understand what XSS or SQL injection are, and they can’t learn. They don’t know why it’s important. They don’t have a technique for checking if their code actually has those vulnerabilities, other than passing it to themselves recursively and asking that other version of themselves to generate some text that might flag if those vulnerabilities were spotted. But, AIs are famously sycophantic so even recursively using itself, it will generate text to “please” itself and probably write something like “your code is great and I can’t spot any vulnerabilities at all! Congratulations! [Emoji] [Emoji] [Emoji]”

To me what’s wild about it is that it’s completely filled with houses, and the houses seem to all respect the orientation of the nearest street.

You’d think that they’d say “Ok, well in this section we have these two roads coming at a narrow angle, let’s just make this a park”, or something to make the places where the two grids join a little less ugly.

The system prompt discovered in the leak explicitly warns the model: “You are operating UNDERCOVER… Your commit messages… MUST NOT contain ANY Anthropic-internal information. Do not blow your cover.”

This is so incredibly stupid.

You’ve tried security.

You’ve tried security through obscurity.

Now try security through giving instructions to an LLM via a system prompt to not blow its cover.

My guess is that 90% of the growth in browser bloat is to support bloated websites.

These days websites can be games, drawing applications, video players, etc. As a result, browsers have basically become operating systems. In addition, the browsers try to support even the most horribly written websites, but that means more bloat in the browser. Meanwhile faster computers mean that people developing websites are just doing more and more javascript, more and more animation, more and more mouse tracking, etc.

If you have an old device with an old browser, a lot of modern websites are completely unusable. I have an old iPad that’s too old to update, and it’s not actually possible to use browse Github anymore. It just ends up with javascript elements on the page that never finish loading. And Github isn’t some site thrown together by someone vibe-coding their first website or something.

There’s a point in that process where you stop caring if you permanently damage things.

I had a bike where the stem of the seat was somehow permanently welded into the tube. Nothing I did worked. I took it to a bike shop and they said there was nothing they could do without damaging the bike. At some point I just gave up and was willing to sacrifice the bike’s frame just to get those two metal parts apart.

I remember that post, and that only makes it stranger. What is the BBC doing that requires that many monitors?

That’s great. I hope he’s basically able to pick and choose his roles these days, and not work if there’s nothing he’s interested in.

If he was smart about investing it, he probably earned enough to retire and live at a middle-class level for the rest of his life. But, he’s probably not loaded. He wasn’t ever one of the main stars of a show, he was always a side character. Plus, lifestyle inflation is a big problem for actors. Say he was making $400k per year on Hawaii Five-O, he probably wasn’t living a modest $80k/year lifestyle and putting away 80% of his earnings.

I would bet he’s still out looking for work so he can live in a house in LA with a pool up in the hills, not in a 1 bedroom apartment in the city.

The irony is that this spending is mostly needed because the US is causing NATO to collapse.

this is the beginning of the end for the AI bubble

The end of the AI bubble has been beginning for years. The end of the beginning of the end of the bubble might take a few more years.

Basically, yeah. He asked Levar if he’d mind if he (Wil) continued the podcast, and Levar gave him his blessing. I think Wil is focusing a bit more on unknown authors though.

I want to believe this is a world where there’s more than just those 2 options.

They can be, or they can be very dense. It depends what you listen to. Many podcasts are just radio shows that actually do go out over the air, but are also repackaged as podcasts. For those, they tend to keep the information dense. The other kind are the stuff that could never be a radio show because it’s just a few guys chatting for a few hours. But, IMO, those can be valuable too because it’s not being rushed to fit in a certain time slot.

but those awful 80s morning shows were still better than these podcasts.

No way. I’d listen to just about anything rather than slide whistles, fake laughs and sound effects.

My most listened to podcast genres are:

• Fans of a sports team celebrate or commiserate
• Cheerful people look at the world through the lens of economics
• A professional voice actor reads me cool short stories
• A science reporter lets scientists gush about their cool discoveries
• Science myths are investigated and sometimes debunked, in an entertaining way (with citations)
• A storyteller tells true stories about mistakes and what we can learn from them
• BBC Foreign Correspondents talk about the stories behind the stories
• The current world is examined from a perspective of how the media delivers the message

They’re saying “There ought to be no gods other than the one I believe in”, despite the fact that other people believe in other gods. They think that those people are delusional and believe in a god that isn’t there, but that they’re perfectly reasonable to believe in theirs. They think it’s absolutely absurd to think that Lord Vishnu had a flower growing out of his navel which he separated into three parts, creating the earth from one of them. But, they think it’s perfectly reasonable that Elohim created the heavens and the earth in six days.

Not only that, but they don’t even believe that this “Lord Vishnu” exists. It’s not that the Hindus got the story wrong and that he was just standing off to the side while Elohim did the work, they think that Hindus are suckers for thinking that he even exists, and that it’s only their god that exists.

If there’s a presumed need for a deity to exist to explain the world (which is absurd), then why restrict it to just one deity? Many believers throughout time have believed that there are many gods, just that theirs are the strongest. But, modern monotheists somehow believe that it’s a fantasy that other gods exist, but not that theirs exists.

What point is that, laughing your ass off?

Some religious people still have a problem with that, but this explanation seems to work for me.

Me: “Do you believe in Ra, the sun god?”

Them: “No”

Me: “Do you believe in Zeus?”

Them: “No”

Me: “What about Odin, or Quetzacotl, or Shiva?”

Them: “No, I only believe in the one true god who–”

Me: “So, you’re basically almost as much of an Athiest as me. Throughout history there have been many cultures who have believed in their gods. You don’t believe in any of those gods, and neither do I. The only difference is that there’s one god that you believe in that I don’t. You’re 99.9% towards being fully Athiest, you just have one remaining god that you still believe in.”

This also helps when they start giving reasons for why what they believe is real because it’s in their bible. You can ask if they’ve read all the holy books of the Aztecs or the Hindus. Why would their holy book be true and not those other holy books? If we’re going to say something is true because it’s in a holy book, then you also have to believe the books that talk about Thor and Odin. If they start saying that everything around was created by god, again, which god? The Hindus have a story for how their various gods created everything, so do the Egyptians. Basically every religion has that story. It’s also useful to ask them what they’d believe if they’d grown up in India, or in ancient Egypt or in Denmark 1000 years ago since almost everybody gets their religion from their upbringing.