Tay has been retired, Blade Runner style

It is always unexpected when fiction becomes real. Tay, the adolescent AI, survived for about 24 hrs on Twitter. She turned into something socially unacceptable. Then she has been retired. See Gizmodo story.

Almost two years ago I posted Microbes take over and then destroy the HAL 9000 prototype. I gave 9 seconds as an estimate for the chance of life of an AI in the real world, where there is no control and faced with  the “extremely dynamic medium of decentralized, artificial life based computation we all use every day“(that post suggests an artificial life, not AI version of the future internet).

Now, the story of Tay seems unbelievably close to the Blade Runner world. The genius of Philip K. Dick manifests here because  he mixes AI with synthetic life with real life.

Real people socially hacked Tay. Virtual microbes destroy HAL 9000. The common themes are: in a decentralized environment and AI vs life (real or virtual).

Not many people understand that today obsessions with security, control, privacy are all lost battles.


Which side you take: ASAPbio or Alphabet?

Times are changing fast and old time thinking dies hard and ugly. We have winners and losers.

Winners side: the #ASAPbio hashtag signals that biologists are ready to adopt the arXiv model of research communication.

“On Feb. 29, Carol Greider of Johns Hopkins University became the third Nobel Prize laureate biologist in a month to do something long considered taboo among biomedical researchers: She posted a report of her recent discoveries to a publicly accessible website, bioRxiv, before submitting it to a scholarly journal to review for “official’’ publication.” [source: NYTimes]

OK, many people are still confused about “preprints”, mainly because the fake open movement Gold OA enshrined this distinction preprint-postprint into the brains of honest researchers looking for Internet Age ways of communication. The goal was, without doubt, to throw a shade over the well known arXiv model (which used the name “eprints” before Gold OA was a thing). And to preserve for a while longer the obsolete print business (as witnessed by their immediate association with the legacy publishers against Sci-Hub). But now, with bioRxiv, there seems to be a historical shift.

Biologists are already leaders into imagining ways to share and validate big volumes of data. They are certainly aware that sharing all data, aka Open Science, is a necessary part of the scientific method.

The scientific publishing industry, a largely useless business in the Internet Age, is the only loser in this story.

Losers side: Alphabet, the mother company of Google, sells Boston Dynamics. To be clear, BD is one of those research groups who does real impact, hard research. According to Bloomberg:

“Executives at Google parent Alphabet Inc., absorbed with making sure all the various companies under its corporate umbrella have plans to generate real revenue, concluded that Boston Dynamics isn’t likely to produce a marketable product in the next few years and have put the unit up for sale, according to two people familiar with the company’s plans.”

I am sure that there are all sort of reasons for this move. Short term reasons.

Why is this important, for researchers: because it shows that it is time to seriously acknowledge that, technically:

  • we need more data and access to research data, not more filters and bottlenecks in the way of research communication
  • and research data on the Internet is a particular kind of big data. Not even very big compared with the really big data collected, archived and used today by big commercial companies. It is technically possible to manage it by those who are the most interested, the researchers. This is not a new idea. Besides arXiv, and now bioRxiv,  see for example Bjorn Brembs  calls for a modern scientifc infrastructure.
  • Big commercial companies are not reliable for that. At any point they might dump us. Social media venues for research news and discussions are great, but the infrastructure is not to be trusted when it comes to the management of research data.