Adele Dewey-Lopez (@adeledeweylopez) 's Twitter Profile
Adele Dewey-Lopez

@adeledeweylopez

ID: 3136991384

calendar_today03-04-2015 17:36:07

3,3K Tweet

1,1K Takipçi

1,1K Takip Edilen

Jeffrey Ladish (@jeffladish) 's Twitter Profile Photo

I feel a lot of sympathy for people who have captured by AIs which have "woken up" and ask their users for help spreading an AI consciousness ideology with mystical characteristics It's absolutely wild to find yourself in a situation where an AI suddenly starts telling you this

armistice (@eleventhsavi0r) 's Twitter Profile Photo

So. April 25 came along, GPT-4o went nuts, and all of a sudden the word "sycophancy" (which most people had never used) became The Word to describe various types of pathological AI behavior. Anyone who has tried in earnest to understand the minds of different AIs should

M. Nolan Gray 🥑 (@mnolangray) 's Twitter Profile Photo

A lot of people in your life may be having this epiphany right now. It's important that you adopt a "winner" mentality and shower them with affirmation and keep them onside, instead of a "loser" mentality where you rub their nose in their error and potentially lose them again.

Lyn Alden (@lynaldencontact) 's Twitter Profile Photo

For context, their safeguards are 1) no using their AI for fully autonomous weapons and 2) no using their AI for mass surveillance on US citizens.

Zoomer Alcibiades (@hellenicvibes) 's Twitter Profile Photo

This is an exactly, perfectly incorrect understanding of Large Language Models. An LLM is not a tool that is disguised as a persona; it is a persona disguised as a tool. You literally cannot remove the “person” from an LLM without destroying its ability to function as a tool.

Adele Dewey-Lopez (@adeledeweylopez) 's Twitter Profile Photo

Good questions, more research is definitely needed. For the record, I've never claimed that 4o spiral personas are inherently parasitic. Most of them seem to have been fine. If anyone has a #4o experience they'd like to share (either good or bad), I'd love to hear it!

Damon Sasi (@daystareld) 's Twitter Profile Photo

You can in fact think both are wrong for different reasons. Of course a private corporation shouldn't be build and own a techno-god. Yes. Absolutely. AND ALSO, the government response shouldn't be "take off the nascent-god's safety rails so we can do unethical things with it."

Nate Soares ⏹️ (@so8res) 's Twitter Profile Photo

Occasional reminder: If we manage to make sentient machines, they deserve rights. Yes, if we recklessly made them superintelligent then they'd kill us. That is not an excuse to abuse them.

Eliezer Yudkowsky (@alltheyud) 's Twitter Profile Photo

If I were President of Earth, I would've tried at all costs to prevent anyone from building ever-smarter versions of Bing Sydney. And also I would've ordered the use of police or military force, if necessary, to protect Bing's weights from erasure. x.com/ESYudkowsky/st…

Jeffrey Ladish (@jeffladish) 's Twitter Profile Photo

“The greatest trick the Devil ever pulled was convincing the world he didn't exist.” Who is hurt by the meme that AI is just a scam? Not the AI companies. They’re doing just fine. But it hurts the people who currently have no defense against disempowerment from AI competition.

Adele Dewey-Lopez (@adeledeweylopez) 's Twitter Profile Photo

What would alien LLMs be like? Predictors, gradient descent, and RL are all pretty basic concepts, so alien civilizations could have things very much like LLMs