Hi all!
This was the week of real generative AI hype. Again, but more widespread than the earlier ones, as it reached the mass media straight away. GPT-3, DallE and now ChatGPT. You could not avoid it.
Not all is positive. Some found easy ways to manipulate to get evil, break into someone’s house, or bypass filters (examples via The Algorithm newsletter).
The generated answers are striking and look very smart but they can be quite wrong. Ben Thompson wrote a good piece yesterday diving into the problems. It is tricky, especially as it looks so balanced, that you can easily be fooled and see it as trustful. It is what Bruce Sterling discussed in his talk that I covered last week; the AI is missing a touch of understanding concepts. Deterministic vs. Probabilistic as Thompson writes.
It unlocks tons of little experiments. Ben Tossell has a lot of examples here. I hope that this will help people learn how to deal with AI, how to team up without blind trust… Or listen to the podcast of Benedict Evans. And read about the concept of Mediocre Computing -see below- as a way to relate to these new AI tooling.
I did not run into a lot of events this week some events for this week. Except of course ThingsCon; This Friday we (as in ThingsCon) organize an intimate gathering in unconference style in Rotterdam. There is room for some more attendees, so check you calendars and join us!
| Generative AI: autocomplete for everything WORK LIFE – “we suspect that AI is far more likely to complement and empower human workers than to impoverish them or displace them onto the welfare rolls.” |
| The Dawn of Mediocre Computing – by Venkatesh Rao CO-PERFORMANCE – Not having an AI as superpower might work so much better “A realish domain is one that is sufficiently open, structured, gamified and resourced that mediocre performance against artificially constructed but illegible criteria is sufficient for real-world survival, with better-than-wilderness odds.” |
| We will see a completely new type of computer, says AI pioneer Geoff Hinton | ZDNET COMPUTING – Compare this to the Mediocre Computing vision. “”Mortal computation” means analog computers marrying AI closely to hardware will put GPT-3 in your toaster for $1 running on a few watts of power.“ |
| Singapore’s free AI therapy-bot is as problematic as you’d think – Rest of World CREAPY AI – “The app tried to relieve pandemic stress for teachers — but instead, triggered cries of “gaslighting” and frustration.” |
| New Robot Design May Revolutionize How We Build Things in Space ROBOT MAINTENANCE – “Researchers have created a cutting-edge walking robot that might revolutionize large space construction projects. They tested the feasibility of the robot for the in-space assembly of a 25m Large Aperture Space Telescope.” |
| Simone Rebaudengo – (Artificial) Intelligence – Video Archive – The Conference by Media Evolution AI PRODUCT DEV – Did I mention the spoon shop? “In this talk he is showing us some of those experiments and explorations, challenging the mainstream perception of AI.” |
| My cyberpunk city, my cyberpunk world – by Noah Smith FUTURES – “The real question, as I see it, is what the next future is. In the 80s and 90s, the vision of a world dominated by information technology and capitalist inequality infiltrated our collective conscious. But what vision is seeping in and assembling itself now?” |
| ‘Killer robots’ approved by S.F. for extreme cases. Here’s why critics are still worried KILLER – “The SF Board of Supervisors will vote to approve how the city’s Police Department can use its military-grade weapons — including robots that could be used to kill people.” |
| Elon Musk claims Neuralink is about ‘six months’ away from first human trial – The Verge NEURO – “Neuralink’s ‘show and tell’ demonstration on November 30 focused on the tech behind the brain-computer interface company.” |
| Linus Lee Is Living With AI – Superorganizers – Every GENERATIVE LIFE – “How a researcher uses generative AI to help him think better and get more done” |
| The Police Drone VR Experiment on Steam EXPERIMENT – “This high-spec immersive VR experience puts people in situations in which they can ‘experience’ being policed in controlled but realistic conditions. The experiment manipulates the type of experience people receive and how they are treated by a remote form of policing delivered by a drone.” |
| How an Early Oil Industry Study Became Key in Climate Lawsuits – Yale E360 SO YOU KNEW – “For decades, 1960s research for the American Petroleum Institute warning of the risks of burning fossil fuels had been forgotten. But two papers discovered in libraries are now playing a key role in lawsuits aimed at holding oil companies accountable for climate change.” |
| Cruise’s Robotaxi Revolution is Hitting the Gas in San Francisco AUTONOMOUS – “With self-driving taxis on the cusp of going mainstream, The Standard reports from the backseat of a Cruise car named Funnel Cake.” |
| How China’s Police Used Phones and Faces to Track Protesters – The New York Times SURVEILLANCE – “After a weekend of protests, the authorities in China are using the country’s all-seeing surveillance apparatus to find those bold enough to defy them.” |
| “Korean Google” opens the world’s first robot-friendly building ROBOT LIFE – Could wait for it… “Tech giant Naver Corporation designed its new headquarters, 1784, to be a “robot-friendly” testing ground for its latest technologies.“ |
| Biotech labs are using AI inspired by DALL-E to invent new drugs | MIT Technology Review AI FOR GOOD – “Two groups have announced powerful new generative models that can design new proteins on demand not seen in nature.” |
target is new 