October 30, 2019

Computerized fakery intrude onto our reality

Futuristic tweets by @vakibs

The fundamental problem with our computing culture is that we glorify the practice of hiding computational  resources from users. Google wants you to believe that you have infinite storage and computation for search. Of course, it’s nonsense. But it’s a strongly crafted illusion.
https://twitter.com/vakibs/status/1188372503816757248?s=19
This pretense about infinite computational power to relatively dumb users is maintained by Moore’s law. But what a waste it was!
We destroyed the fundamental basis of computer science, which is all about optimizing resources. And made users into idiots, computationally speaking.
The stupidity of hiding computational resources is continuing well into the AI age. In fact, it got far more stupider. It is very easy to be a dumb idiot these days, even if you are a programmer. 
https://t.co/jD9c4lFe2l

"We are embedding some very old and disparaging notions about human beings and their place in the natural order into our future technological infrastructure. Engineers at leading tech firms and universities tend to see human beings as the problem and technology as the solution."

If we see the researcher ecosystem as an organism, it's as if this organism pays all attention to the sympathetic nervous system (triggered by novelty, surprise, fear etc) and no attention at all to the parasympathetic nervous system (responsible for homeostasis, growth, health).
https://twitter.com/vakibs/status/1189500640298446849?s=19
The reason why computer science is in this ridiculous situation is a self-afflicted misery. Other scientific disciplines don't behave like this. Physicists are able to come together and build large projects that address serious long-term challenges. Why not CS and AI people?
It is a sad judgement on the field when the only large AI projects are within the confines of private corporations, or  in the secret service departments of governments. The source-code from researchers is fragmented, although they have a high motivation to collaborate and share.

We should completely redesign the idea of a scientific conference. They are all becoming giant Jamborees, producing a ton of trash, and their contribution for scientific culture and connectivity is getting more and more questionable.
https://twitter.com/vakibs/status/1189441040551436288?s=19
Instead of conferences whose ephemeral existence is erased and forgotten in less than a year, I think we should promote scientific pilgrimages. That allows a steady deposition of scientific culture at a geographical place. This promotes “presence” - sense of really being there.
In fact, the model of a pilgrimage is how scientific culture was nurtured for the most of our history. Definitely in India for millennia, until the pilgrimage routes were disrupted in the Middle Ages. This was also the model for Europe, especially in the renaissance era.
We still have relics of this practice. A postdoc will spend some time in a foreign country to imbibe the scientific culture there, and contribute to that. My guess is that most scientific output is  sited through such visits. It’s a pity we don’t do that for senior researchers.
In the current era of instantaneous communication and almost infinite archival of knowledge, the journeys in a scientific pilgrimage can be chronicled and shared for the wider audience. Physical presence needs to be used to maximize cognitive presence. That is what is needed!
When a senior researcher takes time off during a sabbatical year, the utility of that time can be maximized by adhering to a specific pilgrimage route. When multiple researchers do that, the likelihood that they will meet each other and inspire each other will increase.

The only sensible comparison with this Data=Oil model is that if we don't take adequate care against such a development, capitalist exploitation of both data and oil will lead to shitty monopolies that will threaten democracy.
https://twitter.com/vakibs/status/1189540901204496385?s=19

No comments:

Post a Comment