Dan Ryan

Mathematician and Data Scientist

[Dan Ryan @drdanielryan]

Had a chat with Gemini about my view on AI Doom. I think it gets it quite well. Much more than many people building the tech https://t.co/ca2Q0MKFiS

[Liron Shapira @liron]

I’m appalled to see my esteemed AI safety friends shrink from the battlefield of public comms like scared chickenshits.

Today the Center for AI Safety (@cais) abruptly fired John Sherman (@ForHumanityPod) from his new role as Director of Public Engagement, in the wake of a Nirit Weiss-Blatt tweet:

John is a Peabody award winning journalist and host of For Humanity, the most popular show about high AI extinction risk (177k subscribers). CAIS hired him because they understand perfectly well that it's crunch time — time to fill in the “missing mood” around our P(doom). Time to show some emotion!

But suddenly today, a clip surfaced… a clip of John remarking on his show that people would grab their pitchforks and burn down the AI labs if they knew the level of extinction risk these companies were subjecting humanity to.

Nirit Weiss-Blatt tweeted it, and Marc “Calling-for-AI-treaties-means-calling-for-violent-airstrikes” Andreessen promptly retweeted it.

And just like that, CAIS folded like a napkin at an origami convention. Their account responded to Nirit's tweet saying yep, you got us, John needs to go because:
> The connotation of statements like this do not reflect CAIS's values.

What? Have you listened to a single For Humanity episode?

John’s connotation is that of a concerned dad wondering why people aren’t rising up against those who would end the world — those reckless tech cowboys who gamble away our future because they think building superintelligent AI is just the next step after building AI for showing Instagram ads.

The missing mood that everyone else in our tribe studiously avoids, because god forbid the reckless fools holding the future-ruining powerball turn around and label them “extremists” instead of squeaky clean neutral parties that everyone always holds in good standing. YOU RECRUITED HIM BECAUSE YOU REALIZED THIS.

To be clear, John's pitchfork statement is a totally reasonable hypothetical about people with a sufficiently high P(doom) from AI. It's like noticing gain-of-function research in the Wuhan lab and saying, "If people knew what was happening here, they'd burn this place down!" Turn on the cable news and you’ll see on a regular basis pundits hypothesizing about when to bring out the pitchforks. The implied logic of this common type of hyperbole is that if a malicious/misguided actor is doing something so wrong that the majority of us would burn it down, then it’s even easier to stop it using rule of law once people get on the same page.

But Nirit had a second bombshell quote ready to seal John’s fate…
John: “I think that our best chance is that a 'warning shot' that is catastrophic kills a few million people, but not any more than that, several hundred thousand, or whatever it is, and everyone in the world clearly knows that it was the AI companies that are responsible.”

HE’S TALKING ABOUT KILLING PEOPLE! WE GOT HIM FOLKS!

Wait, hold on… didn’t New York Times opinion columnist Ross Douthat just say this in an interview last week?
“In an odd way… if you want regulation, you want limits on AI, maybe you should be rooting for a scenario where some version of hallucination happens and causes a disaster, right?”

Better watch your back, Ross!

Just kidding — even the New York Times doesn’t cancel people like that anymore; the EA/rationalist/AI x-risk community has sadly taken this kind of paranoid self-policing to new heights.

Allowing Marc Andreessen to effortlessly pull the rug out from under John Sherman isn’t an isolated incident or specific to CAIS. I’m constantly cringing, watching our plucky community of AI safety organizations simultaneously declare “PR is our highest leverage focus area!” and “NOOO WE CAN’T DARE TO SAY THIS, TONE IT DOWN!!!”

So please, if you’re someone who understands the urgency of AI x-risk, and you’re thinking about walking back a perfectly reasonable statement, hit me up for moral support. Or a pep talk about how to not be chickenshit.

[Nate Soares ⏹️ @So8res]

I wrote a book with Eliezer Yudkowsky. It’s about how smarter-than-human AI is on track to kill us all, written in hopes of bringing the conversation into the mainstream. Lots of people are alarmed; ~nobody wants to sound alarmist. The time is ripe. Preorders help. Link below.

[Dan Ryan @drdanielryan]

@Liv_Boeree it would be awesome if you could get Eliezer @ESYudkowsky or Nate Soares @So8res on the Win Win podcast to discuss the new book they announced https://t.co/TOWUpJ4PgL

[Max Tegmark @tegmark]

Great discussion about smarter-than-human AGI: Can we control it? Should we race to it? Do we need international collaboration? @demishassabis @Yoshua_Bengio @dawnsongtweets Yaqin Zhang

[PyData Miami @PyDataMiami]

Early Bird registration is now open for PyData Miami 2020. Please take this opportunity to book your spot here: https://t.co/XdKHSHLXUu #pydata #DataScientists #DataEngineers

[Dan Ryan @drdanielryan]

@Navient Can't believe your site is blocking account login the weekend before taxes are due, wtf!

[Dan Ryan @drdanielryan]

@mat_kelcey Maybe @nvidia @rapidsai could help?

[Dan Ryan @drdanielryan]

@iskander That’s an awesome idea. Kind of like multi resolution optimization, e.g. Hyperband.

[Dan Ryan @drdanielryan]

Highly recommend recent @Rspodcast interviewing @robertwiblin about work at @80000Hours. Two of my favorite podcast hosts having an extremely thoughtful and nuanced conversation. https://t.co/ivmdF01Tud

[Demis Hassabis @demishassabis]

1/3 Really proud to share #AlphaStar with the world - a milestone for game-playing AI, and the first system to beat a top pro at #StarCraft2. Massive thanks to @Blizzard_Ent, @Liquid_MaNa & @LiquidTLO for being part of this great journey!

Blog + Replays: https://t.co/6VBDfiGQsu

[Dan Ryan @drdanielryan]

Would love to hear what @ESYudkowsky thinks of @FHIOxford technical report "Reframing Superintelligence: Comprehensive AI Services as General Intelligence"

[Dan Ryan @drdanielryan]

@AndrewYangVFA gave a great interview on the @Freakonomics Radio podcast recently. If you don't want to read his whole book "The War on Normal People", I think this is the best way to understand him and his ideas. But, you should read the book, it is great.

[Dan Ryan @drdanielryan]

Scientists identify vast underground ecosystem containing billions of micro-organisms https://t.co/AMDDQedN1b

[Murray Shanahan @mpshanahan]

Hinton: "if you want to get a paper published in [ML] now it's got to have a table in it ... datasets ... methods ... and your method has to look like the best one. ... I don't think that's encouraging people to think about radically new ideas" - https://t.co/tjbErVdjxe

[Dan Ryan @drdanielryan]

@SpencrGreenberg just finished the Primals survey and scored close to zero on “Alive”. How do you think that dimension relates to rationality and a naturalistic world view?

[Dan Ryan @drdanielryan]

Loving “The Nashville Session 2” from @NewMastersounds! Going to be grooving to this on a daily basis for the foreseeable future.

[Dan Rather @DanRather]

When someone starts an argument with "I'm not a scientist, but..." maybe we should stop listening to them weighing in on science. And maybe news shows should stop asking these pundits to talk about something they’re not qualified to talk about.

[Dan Ryan @drdanielryan]

Loved this essay by @Zat_Rana. He has many more that are great too. Check him out if you are into philosophy, psychology, cognitive science. https://t.co/D77nCnkEhI

[Rob Reid @Rob_Reid]

Amazing work at Stanford could reverse blindness. And one day, enable profound neural augmentation (!) Learn about the astounding neuroscience and engineering behind all this from its inventor, right here: https://t.co/00B2pHm6Cb