Zvi Mowshowitz
@TheZvi
Blogger world modeling, now mostly AI and AI x-risk, at Don't Worry About the Vase (http://thezvi.URL.com on SS/WP, LW), founding Balsa Research to fix policy.
Bumping this now that we've had a few days to try out ChatGPT Agent. Are we doing the thing? (I have not, so far, found a Thing for it to try and Do.)
ChatGPT Agent reaction thread. Can it do the thing?
For those looking to understand the issue I found this excellent and I agree with most of it.
I've written a new post about "ChatGPT psychosis". Includes a detailed timeline of events leading up to the latest incident with @GeoffLewisOrg Link below.
*once again looks over at the Days Without a Crashout sign on my desk and grits teeth while getting back to writing up the generalized reasoning model that speaks in broken English and got IMO gold, I'm having a totally great day how are you?*
New paper & surprising result. LLMs transmit traits to other models via hidden signals in data. Datasets consisting only of 3-digit numbers can transmit a love for owls, or evil tendencies. 🧵
Okay, definitely not otherwise taking any capital gains for the next three years just in case. I wonder who else will do likewise.
🚨 TRUMP: THINKING ABOUT NO TAX ON CAPITAL GAINS, HOUSES
A preschool (this is real) requires that all students get checked for lice before coming the first day. You do this by taking them to a place people check kids for lice. Does this make the risk of lice on that first day:
This, but switch out utilitarianism, and instead of 'reduce suffering' it's things like 'preventing all the humans from dying' or 'be nice to each other for a change.'
I see the Shrimp Welfare debate has got us doing this version of utilitarianism discourse this morning.
If anything, I worry that thought experiments train people to expect counterintuitive or difficult answers, because only those get written up. Philosophy courses ought to *start with* "Is it okay to kill a guy for a bagel? No."
Polymarket has acquired QCEX, a CFTC-regulated exchange and clearinghouse, for $112 million. This paves the way for us to welcome American traders again. I've waited a long time to say this: Polymarket is coming home 🇺🇸🦅
Debating how I feel about this - I was going to cover the IMO result tomorrow but if this is true then there's a strong argument to wait until Google's results are verified so as not to reward defection. Thoughts?
🚨 According to a friend, the IMO asked AI companies not to steal the spotlight from kids and to wait a week after the closing ceremony to announce results. OpenAI announced the results BEFORE the closing ceremony. According to a Coordinator on Problem 6, the one problem OpenAI…
...and back to work on building it, then, I guess.
At times, AI existential dread is overwhelming
Was playing around with the new agent feature and used this prompt just to see what would happen. I promise I did not write the part that’s circled, it gave that command on my behalf 😳
I'd like to see them try.
We’re going to make Baby Grok @xAI, an app dedicated to kid-friendly content