Discover and read the best of Twitter Threads about #AISafety
Most recents (7)
🧵1/ Just attended a 🔥 @LangChainAI webinar on AI agents, ft. some of the brightest minds in the space! Let's unpack the key takeaways & explore the cutting-edge work being done.
Guests:
@charles_irl
@ShunyuYao12
@mbusigin
@yoheinakajima
@hwchase17
Guests:
@charles_irl
@ShunyuYao12
@mbusigin
@yoheinakajima
@hwchase17
@LangChainAI @charles_irl @ShunyuYao12 @mbusigin @yoheinakajima @hwchase17 🧠2/ Shunyu introduced the core idea of his #ReAct paper, which adds a "thought" between Action & Observation. The open question is, do we need a strict pattern of Thought-Action-Observation or should we just add thoughts as a special type of action, offering more flexibility?
@LangChainAI @charles_irl @ShunyuYao12 @mbusigin @yoheinakajima @hwchase17 🤖3/ Yohei built babyAGI, inspired by #HustleGPT. He used GPT-4 to create an autonomous Founder Bot, even penning a "scientific paper" about it. BabyAGI focuses on completing, generating, and prioritizing tasks. #BabyAGI
1/6 🚨ALERT: Samsung Semiconductor data leak highlights the risks of AI in the workplace 🤖💻 According to recent reports, confidential data from Samsung was inadvertently leaked via ChatGPT, an AI writer used to help fix problems with their source code. #AIrisks #digitalsecurity
This isn't the first time AI technology has raised concerns about security breaches and job replacement risks. Our latest article explains how to use AI technology responsibly #AIsafety #workplacerisks
3/6: Since ChatGPT retains user input data to further train itself, these trade secrets from Samsung are now effectively in the hands of OpenAI, the company behind the AI service. #dataprotection #chatbots
If you react to @timnitGebru and @xriskology that their descriptions of effective altruism are stereotyped and unreasonable, consider:
None of this is new.
———
In 2019, @glenweyl criticised EA in an interview with @80000Hours and went on to post: radicalxchange.org/media/blog/201…
None of this is new.
———
In 2019, @glenweyl criticised EA in an interview with @80000Hours and went on to post: radicalxchange.org/media/blog/201…
In 2021, Scott Alexander wrote “Contra Weyl on Technocracy” and what followed was a series of debates where everyone seemed to talk past each other and left feeling smug.
Track back some of those posts here:
google.com/search?q=glen+…
Track back some of those posts here:
google.com/search?q=glen+…
Weyl: Here is my holistic emotional sense of what’s going wrong with these EAs who are invading spaces in the Bay Area, and I want them to stop.
EAs: Politics alert. Such imprecise and intense judgements. EAs donate to GiveDirectly too – did you consider that?
Sounds familiar?
EAs: Politics alert. Such imprecise and intense judgements. EAs donate to GiveDirectly too – did you consider that?
Sounds familiar?
"Long-termism is not only influenced by far-right eugenics but, in turn, is influencing the billionaires involved in the biggest technology companies in the world including Facebook, Google, Amazon, Microsoft – and the Biden administration."
"Another long-termist,...Nick Beckstead, claims that saving the lives of more “economically productive” rich people is a greater priority than saving unproductive poor people, because this would...increase chances of the desired future of massive space colonisation."
"This will, Bostrom and Shulman argue...“increase world human capital, &,...possibly create individuals with unprecedented levels of cognitive capacity.” They therefore recommend that policy-makers explore “appropriate regulatory frameworks”, & even “a common set of global rules”
And here I was talking about Anthropic AI getting 580M from the ppl behind crypto fruad dumpsterfire everyone is seeing.
"$580M round led by Sam Bankman-Fried,...participation from Caroline Ellison, Jim McClave, Nishad Singh, Jaan Tallinn & the Center for Emerging Risk Research"
"$580M round led by Sam Bankman-Fried,...participation from Caroline Ellison, Jim McClave, Nishad Singh, Jaan Tallinn & the Center for Emerging Risk Research"
Never heard of the Center for Emerging Risk Research but they sound like they're in the cult.
actuia.com/english/anthro…
actuia.com/english/anthro…
Anyways, some of the "moral" philosophers at Anthropic are the most ardent desciples of this cult.
When Deafblind activists told them to stop comparing them to LLMs, they said no I won't stop just because you told me to. #AISafety.
When Deafblind activists told them to stop comparing them to LLMs, they said no I won't stop just because you told me to. #AISafety.
Wow - in 8 tweets I just learned and un-learned more about the mysteries of deep neural networks than I've probably learned or un-learned about them in the last two years.
This is the start of something really really big... also a huge door opened for federated learning.
This is the start of something really really big... also a huge door opened for federated learning.
This technique really seems to get a foothold on managing the intelligence in an AI model. Imagine training 10,000 small models on 10,000 different topic areas and being able to decide exactly what collections of specialties a model was to have.
Heads up #AISafety community!
Heads up #AISafety community!
Also - the natural next question.... can it be reversed? Can I un-merge two models that have been merged?
Example: If I trained an MNIST model on half the digits in MNIST, could I then use it to remove the intelligence of those digits from a normal MNIST model?
Example: If I trained an MNIST model on half the digits in MNIST, could I then use it to remove the intelligence of those digits from a normal MNIST model?
Today, at @EPFL_en @EPFL @EPFLevents, @harari_yuval, arguably today's most influential thinker, is giving a talk to an audience of 2700 (!).
Harari wrote the highly celebrated books Sapiens, Homo Deus, and 21 Lessons for the 21st Century. I highly recommend them!
Harari wrote the highly celebrated books Sapiens, Homo Deus, and 21 Lessons for the 21st Century. I highly recommend them!
He started with such a major claim: humans are hackable.
Using Biology, Computing and Data, science delivered the Ability to Hack Humans.
For the better, and the worse. If not for the worst.
Technology is upsetting our world. We urgently need to set it in the right direction.
Using Biology, Computing and Data, science delivered the Ability to Hack Humans.
For the better, and the worse. If not for the worst.
Technology is upsetting our world. We urgently need to set it in the right direction.
Unfortunately, Harari claims, today's politics is almost solely about nostalgic fantasies. They discuss issues of the past.
He argues that we should instead focus on issues of the future. In particular on algorithms and data, who are bound to play an increasingly critical role.
He argues that we should instead focus on issues of the future. In particular on algorithms and data, who are bound to play an increasingly critical role.
