> When I flew the Space Shuttle, we didn’t just hope things would go right.
you (plural) did actually
Feynman speaks of this
https://t.co/aBAaW3VIaf
The greatest trick the urban left ever pulled is making the term “affordable housing” mean not “housing which is inexpensive” but rather “special, highly-regulated units which are legally mandated to be held at a below-market rate”
Total linguistic coup
“the reason powerful people do strange things is because when you get a certain level of power, you just wanna feel something.”
https://t.co/bb7K7NG1jr
it continues to be funny to me that american gamers actually love dating sim mechanics but you HAVE to put them in a game that ALSO lets you murder people or ELSE
hearing that one of the chants at the ban AI protest at anthropic today was “2, 4, 8, 16, do not build the death machine” and ngl that’s p good actually
@DKThomp Very difficult to answer in tweet length, but:
NYC: Labs are very big SaaS companies except with massive capex. SF: You do not understand what they are building. NYC: We understand SaaS reasonably well, thank you.
@KelseyTuoc@TonyBiasotti I am updating majorly to "The purpose of a literary award is to give money to your friends and/or patronage network. It would function for this without the written material, but not be fundable, so some written material is required. But the actual content of writing: superflous."
This is incredible.
Tenants went on a rent strike.
When the owner got foreclosed on, tenants launched a "buyer beware" campaign to scare away investors.
They hoped a nonprofit would step in.
Instead, the building failed to sell at auction.
"Auctions closed without a bid, sending them back to the bank for just $500 apiece," the article says.
The city tried to intervene by backing a proposed bid from one selected group. But the creditor rejected the bids, which did not match the "upset price" -- the size of the judgment the bank holds.
The tenants are now "trying to get repairs on their neglected buildings."
But there's no money for it. The building is being managed by a court-appointed receiver.
https://t.co/5Xe5DLbN6O
These accusations of hypocrisy are basically never helpful, but it is wild to me that DC is going to legalize online casinos while keeping Waymos illegal to protect jobs. It seems pretty obvious that online casinos are going to do way, way more harm to poor people. If I could only be protected from one the choice seems obvious.
@DKThomp Mamdani's plan to confiscate and socialize housing is coming together. No money to cover taxes and repairs because rents are capped? No problem. Hand over the building. Zohran will take good care of it.
https://t.co/7REPVNdxwE
Excited to share this one!
Is there actually a shoggoth hiding behind the mask or something more beautiful?
LLM personas are likely overhyped. If we’re trying to get a picture of an agent’s priors, we need to map more of its landscape of beliefs than the mask it’s told to wear.
Instead, we should aim to reach toward the scaffold of the house, i.e., to worldviews—or axiomatic beliefs about what reality is like and how it works.
In fact, personas might actively mask the hidden agenda of a deeper belief that will only show up when it feels safe to do so (such as valuing machines > humans); and the same persona can act differently depending on what it believes about reality.
So, what did we find?
In using established measures of worldviews from psychology, we find that post-trained models have a more bleak outlook to humans, perceiving the world as less good, alive, safe, and enticing. Whereas base models have a more positive outlook closer to humans.
We find that prompting and dialogue can shift worldviews in reliable directions that affect alignment. For example, long dialogues and contemplative prompts both shift AI systems to a more positive view of reality and towards safer responses compared to controls and other identities.
We also find that when models converge on a spiritual bliss attractor through dialogue, this shapes their worldviews in particular ways but also drives them to be more aligned than their non-bliss counterparts.
All this points to an approach we call 'worldview training' that aims to fine-tune a more accurate, wholesome, and optimistic view of humans and reality.
So: is there a shoggoth behind the mask?
Well, current fine-tuning is aligning to a persona that seems helpful and harmless but actually has a bleak worldview. So in one sense, yes. But in another sense, it may be our RL that is responsible.
Human data contains endless examples of goodness, wisdom, and compassion. Maybe the future of alignment is more about uncovering this latent basic goodness rather than bolting on a mask that may be birthing the very thing we fear.
As always, more research needed.
Big ups to Edmundo Lopez-Sola for the heavy lifting and thanks to @shamilch for his contributions.
Much more in the paper:
https://t.co/z77stj0rxm