Saturday, September 12, 2026

Neil deGrasse Tyson And Jaron Lanier on the AI Illusion



In "One-Half a Manifesto", Lanier criticizes the claims made by writers such as Ray Kurzweil, and opposes the prospect of so-called "cybernetic totalism", which is "a cataclysm brought on when computers become ultra-intelligent masters of matter and life."[22] Lanier's position is that humans may not be considered to be biological computers, i.e., they may not be compared to digital computers in any proper sense, and it is very unlikely that humans could be generally replaced by computers easily in a few decades, even economically. While transistor count increases according to Moore's law, overall performance rises only very slowly. According to Lanier, this is because human productivity in developing software increases only slightly, and software becomes more bloated and remains as error-prone as it ever was. "Simply put, software just won't allow it. Code can't keep up with processing power now, and it never will.


 Lanier criticized the sometimes-claimed omniscience of collective wisdom (including examples such as the Wikipedia article about him, which he said recurrently exaggerates his film-directing work), describing it as "digital Maoism".[25] He writes: "If we start to believe that the Internet itself is an entity that has something to say, we're devaluing those people [creating the content] and making ourselves into idiots."[25]

His criticism aims at several targets that concern him and are at different levels of abstraction:

any attempt to create one final authoritative bottleneck that channels the knowledge onto society is wrong, regardless of whether it is a Wikipedia or any algorithmically created system producing meta information,
it creates a false sense of authority behind the information,
a sterile style of wiki writing is undesirable because:

it removes the touch with the real author of original information, it filters the subtlety of the opinions of the author, essential information (for example, the graphical context of original sources) is lost,

collective authorship tends to produce or align to mainstream or organizational beliefs,
he worries that collectively created works may be manipulated behind the scenes by anonymous groups of editors who bear no visible responsibility,

and that this kind of activity might create future totalitarian systems as these are basically grounded on misbehaved collectives that oppress individuals.

n December 2006, Lanier followed up his critique of the collective wisdom with an article in Edge entitled "Beware the Online Collective".[27] Lanier writes:

I wonder if some aspect of human nature evolved in the context of competing packs. We might be genetically wired to be vulnerable to the lure of the mob... What's to stop an online mass of anonymous but connected people from suddenly turning into a mean mob, just like masses of people have time and time again in the history of every human culture? It's amazing that details in the design of online software can bring out such varied potentials in human behavior. It's time to think about that power on a moral basis.

Lanier argues that the search for deeper information in any area sooner or later requires finding information that has been produced by a single person, or a few devoted individuals: "You have to have a chance to sense personality in order for language to have its full meaning."[27] That is, he sees limitations in the utility of an encyclopedia produced by only partially interested third parties as a form of communication.


https://en.wikipedia.org/wiki/Jaron_Lanier

No comments:

Post a Comment