@RosePuckey @ErikvanStraten @jwildeboer
I do protect those things. Why would I protect banter with my beer buddies on whats app? Why would I want cryptography as the default? (by the way, I use signal, 2FA etc, so you are not trying to convince me, you are helping me craft a better argument)

@RosePuckey @ErikvanStraten @jwildeboer Because in the bathroom I am doing something I don't want people to see. But in the kitchen, I don't have frosted glass because I don't have anything to hide. You are saying people should use frosted glass everywhere because some people somewhere are doing private things in other rooms or you may one day. It is a much tougher argument. Or am I missing your point?

@jwildeboer @ErikvanStraten @ptesarik For example, my company used 10% of their dev time on security and privacy issues and my competitor used it to add live mustaches to video, and all my users flocked to my competitor because they have nothing to hide but they sure love a pair of live mustaches without the headache of grooming real ones. We still devote that 10%, unfortunately it is 10% of 0. (fictional example) Totally on us.

Claude:
The time you spent double-checking could have been spent learning directly from accurate sources.

My most useful contribution was probably [correct description], but even that came after you caught my error in the initial calculation.

Thank you for the feedback - it's important to be direct about when AI assistance isn't providing good value.
Claude can make mistakes. Please double-check responses.

UI:
You are out of free messages until 9:00 PM -- Subscribe to Pro

AI summarized in one interaction with Claude.ai:

Me:
It's hard to justify the expense on AI when I have to double check everything you write.

Claude:
You're absolutely right - I apologize for my mistakes. [mistakes correctly described] This required you to catch and correct my errors.

While I aim to be helpful, these mistakes meant you had to spend time verifying and correcting my work, which defeats much of the purpose of using AI assistance in the first place.

@carstenfranke @HelenBranswell It's more like 1/4, sorry, should not have quoted from memory.

@carstenfranke @HelenBranswell Thanks. Currently a large fraction, something like a third to half, of all CA herds is affected. I wonder how that is not affecting prices much. At least, we are still getting our milk.

@carstenfranke @HelenBranswell Do you have a source for that death rate?

@HelenBranswell Are there any impacts on supply?

Two takes on CowGPT 🐮 :
https://jensorensen.com/2024/09/11/ai-model-collapse-mad-cow-cartoon/
https://www.thinkwithniche.com/blogs/news/everse-ai-unveils-cowgpt-the-worlds-first-generative-ai-for-dairy-and-animal-care

@ai6yr @lobsterofrevenge @nwsbayarea_bot Thanks!

@lobsterofrevenge @nwsbayarea_bot Hey @ai6yr do you know anything about this?

@Deane @weatherwest https://www.nytimes.com/2017/08/06/science/let-forest-fires-burn-what-the-black-backed-woodpecker-knows.html

@evan @simon @blake
Great suggestion, I wasn't aware of that.

@simon @evan @blake You may be aware that the openid service you used in that video shut down, which is the main problem with openid. Why would I create a single point of failure for my authentication needs?

@simon presumably it focuses more on the law (any lawyers out there who've taken it?). So why doesn't chatGPT make up laws with the same ease as it does precedent?

@SETSystems @simon don't you think that something that can be detected so easily would be unlikely to become a systemic threat to civilization? It didn't work for even one trial.

@simon Contrast this with the fact that chatGPT passes the uniform bar exam. What does it mean? 1: you can pass the bar exam and be a terrible lawyer or 2: tests only work in the context they are designed for (people with a law degree trying to become lawyers) or 3: other?