Emma Pierson, a UC Berkeley computer scientist with a BRCA mutation and family history of cancer, uses her Atlantic essay to weigh personal medical upside from faster AI against wider downsides including unemployment, surveillance, autonomous weapons, and eroded human purpose, ultimately favoring deliberate slowdown of generalist systems even if it postpones breakthroughs like new cancer therapies.
New piece in @TheAtlantic!
We always hear that AI will cure cancer, and I would immediately benefit if it did. But I argue that racing ahead on generalist AI models creates unclear benefits for cancer that are outweighed by broader societal harms.
(Gift link in next post)
Pierson questions whether superhuman AI will quickly convert intelligence gains into real-world cures, noting constraints on biological data, ethical experiment rules, and the gap between domains like chess and messy human biology where tools such as AlphaFold have yet to trigger drug revolutions.
The short life of Anthropic’s Claude Fable 5, released with biology-query blocks then fully suspended after export controls, is presented as evidence that current safeguards and governance cannot keep pace with rapid capability jumps.
Many users condemned essays arguing that AI cancer cures fail to offset societal harms as insensitive or disgusting toward patients, while others urged accelerating progress as humanity's best hope against death and risks.
No Digg Deeper questions have been answered for this story yet.
@tszzl But I have family members with cancer, and their lives are worth more to me than some half baked hypothetical invented by someone who can’t even predict what the next week will look like. Please do not pause, people’s lives are on the line. I beg you, don’t listen to these people
Gift link: https://www.theatlantic.com/technology/2026/06/ai-cancer-progress/687654/?gift=yDcMMOSj-3oJk8475KVVBne5CfZXWcmKJJrUAGArmLA&utm_source=copy-link&utm_medium=social&utm_campaign=share
@2plus2make5 @TheAtlantic This is extremely disgusting and disrespectful to actual people with cancer. Do you not any loved ones with it?
@tszzl but in 100% of cases, some segment of people get scared or upset and demand that things stop changing
@2plus2make5 @TheAtlantic What the fuck did I just read? Your priorities are fucked up beyond hope. You're ranking your own personal enjoyment higher than the death and suffering of tens of millions of cancer patients. Reflect on your article the next time you attend a funeral.
@tszzl you agree with this article????? it was super weak imo. dowsnt make the top level claim wrong but be careful what ur promoting
@tszzl "there are many cases where you’d want technology to slow down even at great cost of life"
-doomers
@tszzl There are many hypothetical cases in which this is true. I don't think it's true in reality, at least for me personally.
@tszzl Are you okay?
@IanBaer @tszzl Curing cancer is itself a hypothetical. It's all hypothetical.
We don't know if curing cancer is more or less likely than collapsing modern society.
@2plus2make5 @TheAtlantic what an evil person... makes me want to even more strongly support AI progress...
@tszzl @robbensinger one thing is certain: WE ALL ARE GOING TO DIE IF WE DO NOT WORK ON ASI. asi is probably our only hope that will keep us from dying.
not you too T_T It's not good to condone their talking points, you're fueling their anxiety for speculative reasons. Fire alarms serve a purpose and should not be pulled because someone was smoking in the bathroom There are reasons to use regulations wisely, or dare say even pause however those reasons must be articulated and proven thoroughly unless you want some lab to decide that GPT-2 to too dangerous because the sky shall fall yet again. (yes fable size jumps does have some thorns and issues;; yes some regulation is nec at somepoint esp for certain capability thresholds... but it is crazy to think todays models are going to be looked back as "Very capable" in 10~15 years or less...)
@tszzl Disappointed to see you agree with this, roon
@MaxEllison2048 @tszzl “Curing cancer” can be defined and a path towards it can be sketched out.
“Collapsing society” is an *ADVERTISING SLOGAN* without any real meaning, meant to convince people to buy “ai safety” consulting services. It is nothing it should not even be considered.
@tszzl Am I just totally off base that one of these things (meaning LLM’s) seem to be growing anything like a will or persistent identity? In a way that’s its own danger but even when I had Fable it didn’t feel like a mind that could even “want” anything per se.
@tszzl These things basically aren't even connected. An ML training regime for hypothesizing cancer treatments has almost zero overlap with impersonator models pretending to be human.
@arthrosis_anon @tszzl I don’t care how it is fixed only that it is. Death is the problem and we should be hammering it with everything we’ve got, not acquiescing to the whims of the most neurotic morons on planet earth.
@tszzl Anti-AI sentiment borders on being a suicide cult that forces humanity down with it at this point
@tszzl such as?
