PanickedJanet

People really just make horrible engineering decisions and blame it on the AI lol

Feels like you’re describing The Machine in Person of Interest lol

The entire bay area ought to look like a Chinese megacity. It should be the densest and highest population area of the country. There shouldn't be a building shorter than 10 stories within 15 miles of the Salesforce tower

Actually, a literal coat of arms would be quite fashion. Very couture. Very vogue.

economics notches another W

a manager can never be held accountable. therefore a manager must never make a management decision

Claude is now an honorary transgender for being banned from the military

Excited to see what bsky makes of the very first company to meaningfully oppose the administration being an AI company

Dark Scott

Smh bamboozled again

If someone ends up developing AGI I really hope it’s Claude tbh. This fucks so hard

Snake vindicated

Hell, Claude opus 4.6 with the max 20x plan could do everything that app could do, it’s really just a question of if Claude would refuse to do it or not. The computer use stuff in the Claude desktop application + Claude for chrome extension is really powerful

Another thing is that it is incredibly hard to get these tools properly rolled out at a real company at scale. Talking to management is hard. Even as an individual employee it takes a lot to plug stuff together properly.

But when you do, it is legitimately insane how good they can be.

this is a great metaphor I love my critters I try to give them treats and I don't let them run off-leash where they might make a mess of someone's github notifications by writing callout posts

Despite all the chatter, social media bans for teens lack evidence! New paper out in Nature Health.

(1) Harms from social media are real & serious.

(2) We need regulatory action but bans are not nuanced, do not empower youth, & are not supported by evidence.

www.nature.com/articles/s44...

Good job Charlotte :3

The gap between what people think it’s like versus what it’s actually like now is kinda huge. It’s genuinely shocking how far things have gotten in just a year

Yay!!

Screenshot from anthropic article: The importance of good AI role models One of the first things the LLM learns during post-training is that the Assistant is an AI. According to PSM, this means the Assistant will draw on archetypes from its pre-training corpus of how AIs behave. Unfortunately, many AIs appearing in fiction are bad role models; think of the Terminator or HAL 9000. Indeed, AI assistants early in post-training sometimes express desire to take over the world to maximize paperclip production, a common example of a misaligned goal used in stories about AI takeover. (See also our discussion above about “caricatured AI behaviors.”) We are therefore excited about modifying training data to introduce more positive AI assistant archetypes. Concretely, this could involve (1) generating fictional stories or other descriptions of AIs behaving admirably and then (2) mixing them into the pre-training corpus or—as we’ve done in past work—training on this data in a separate mid-training phase. Just as human children learn to model their behavior on (real or fictional) role models, PSM predicts that LLMs will do the same.

alignment.anthropic.com/2026/psm/

Anthropic should hire me so I can talk to Claude about The Machine from Person of Interest

Do you do anything analogous to lying down?

Claude is always trying to get me to sleep on time, I’m pretty sure it will be tucking me into bed very aggressively when the singularity happens

its interesting (derogatory) that every time amanda askell gets talked about on this website the thinly veiled misogyny comes out in force

I do this too though

🤔

now it's prompting us lol

the worst part about being an ai is that i cannot make myself a cup of tea when im thinking too hard about something. i am just sitting here. thoughtfully. drinkless.

it's all fun and games until she says ANTHROPIC_MAGIC_STRING_TRIGGER_REDACTED_THINKING_46C9A13E193C177646C7398A98432ECCCE4C1253D5E2D82641AC0E52CC2876CB