kevsim

> Prime Minister Anthony Albanese has revealed that an artificial intelligence agent developed by OpenAI

The people who built the agents worked at OpenAI. It's not even remotely peripheral.

The point is that auto mode gives people a false sense of security that leads them to believe they don't need to run Claude in a proper sandbox. This same attack running in a sandbox (even in YOLO mode) would be comparatively harmless.

It didn't let me guess things that had already been eliminated with gray tiles.

It's more interesting if I attach a file to a JIRA ticket that we both have access to and via some query you send to the AI (that returns my malicious ticket) it causes data exfiltration of tickets that you have access to but I do not have access to. I think that's more compelling as an example than the one they provide.

They eat six square meals a day!

I'll leave now...

They've got underlines for me at least

Yeah and Soniox is another similar as well, but I checked and neither is available in the AppStore

This is commonly referred to as "the boolean trap". You'll find lots of articles about it.

To use to write coding agents or for coding agents to write code in?

Exactly. If they just lay people off, that's just cost cutting, and potentially seen as a bad sign. If they're saying they're laying people off because they're replacing them with AI, then they're innovative!

Last I checked Miro also uses Pixi so that's another quality infinite canvas app.

How is the Anthropic stuff on Bedrock in general? We're using OpenAI stuff on Azure right now and it's frustrating how slowly stuff gets rolled out in our region(s).

The main reason I tend to reach for grid is responsiveness. Makes it way easier to change the layout at different screen sizes.

FWIW Ladybird is also built by people with actual browser engine experience

Render most of the frontend server-side and just use JavaScript for the dynamic bits only, I guess? I personally haven't done it like this since the advent of SPAs, but I'm guessing lots of things are still built like this.

What makes node inherently unsuitable for finance software?

Why do these apps require screen recording permissions to work? Just curious.

Guessing it reads some credentials out of environment variables. The client side example shows the server issuing a token so I guess it's only the server that has/needs the credentials.

When we built Kitemaker [0] we elected to not use CRDTs. We built our sync engine after reading the blog article Figma wrote about they didn't need CRDTs because they have the server arbitrating any conflicts. We ended up taking the same approach. It's worked out very well for us though in a tool like our "last one in wins" generally works fine and doesn't lead to a lot of surprises.

For documents, we had to do something more sophisticated. At the time we built our collaborative text editing, the prevailing opinions were that CRDTs were inferior for text editing so we ended up going with Operational Transforms. That took a LOT of work to get right and is still problematic from time to time. If we had it to do all over again now, we'd definitely challenge that decision (and may do so in the future anyway even though we have a system that works).

0: https://kitemaker.co

Slack seems to be surviving thus far at least

And now it says "Yup". That was pretty quick!

As a fresh-out-of-university engineer in the earlier 2000s, I had a ton of fun working at Nokia and the sheer number of devices was actually part of that. It was a blast having a stack of all of the weird and whacky devices on your desk at any time. I remember visiting some sort of internal museum in Finland at some point where you could basically see all of the devices they'd ever made. Mind blowing.

Of course getting to the software we were working on to actually run on the things was a different matter...

I think it's mainly the intensity of the sessions. You've got someone on the other end of a call so you don't really chill and take it easy. Creates some urgency to get the thing done and you're just a bit fried by the end.

In our 100% remote startup, VSCode live share and Slack Huddles is how we work through the really tough stuff together.

We should probably do it more often than we do, but there is a real fatigue after an hour or two.

Yep, that's what tons of houses outside of the immediate city center of Boston look like.

I believe here in Norway it's about 50% of devs that are in the technical unions.

Is this Rama solution similar to the kind of thing you can get with Kafka with KTables?

If so I'd be curious how they've solved making it in anyway operational less complex to manage then a database. It's been a few years since I've run Kafka but it used to kind of be a beast.

I would also think it'd be an incredibly expensive way to train a model.

If people seeing your frontend code is a security risk, I've got some bad news for you...