Lex Fridmans VR-Interview with Mark Zuckerberg
Lex Fridman released a Interview with Mark Zuckerberg in the Metaverse which comes around more like a tech demo than a podcast or an actual interview.
However, i'm impressed. Zuck hinted at this level of photorealism when they first introduced their Metaverse stuff two years ago or so, but seeing this level of realtime CGI-simulation in VR in a decidedly non-formal interview situation is something else.
The interview seems to show two different stages of Metas Avatar tech. While the intro is showing a full body avatar of Lex Fridman which requires their 'old' specialized capture rig with over 100 cameras for a full body scan which they already used when they introduced Codec Avatar back in 2019, it looks like the interview itself uses their updated avatar tech which is able to use smartphones with depth sensors to generate 3D scans of your face, hence the lack of bodies and the visible artifacts on the neck in the interview.
I mentioned this tech in a post about the papers coming out of Siggraph conference one year ago. Here's the video accompanying the specific paper about those avatars from phone scans:
I’m very sure this interview was heavily prepared and i wonder why they didn’t do the full interview with their complete bodyscans. Keep in mind that this tech is at a prototype stage for the moment and it's not clear if this was rendered on headsets' computing powers or a connected PC in the background. I'd guess the latter.
Also, these highly realistic VR-avatars are only compatible with Metas Quest Pro, which's next model is coming sometime in 2025 and will reportedly include this updated avatar technology. It was clear to me that their cartoonish and cheap looking avatars wouldn't stay that way, but i'm impressed that they got it working from a paper to an impressive Fridman-tech demo within a year.
We'll see how Apples 'Spatial Personas' for Facetime on their VisionPro will hold up to this, and ofcourse, adaption of this tech in the real world is a complete other story. VR might deliver the most kickass photorealistic virtual experiences — if it’s still a hustle to get it running and if there’s cable dangling around your neck, adoption will still be limited to a hardcore enthusiast market.
BullshitjobsGPT and the crisis of agency
In B.S. Jobs and the coming crisis of Meaning for the excellent Summer Special of the New Atlantis magazine focused on artificial intelligence, Brian J. A. Boyd warns of an impending crisis of meaning due to the automation of knowledge work in white-collar jobs.
Many lines have already been written about the destruction of what David Graeber called "bullshit jobs" in his book of the same name, most notably in the New Yorker, when scifi author Ted Chiang compared the automation potential of AI with the optimization process of consulting firms like McKinsey and fitting the bill, just recently Ethan Mollick published a study on the use of ChatGPT-4 at the Boston Consulting Group, which showed that employees using the AI-model were way more productive and worked faster than those who did not use AI.
David Karpf nailed it with this comment on that study: "Hey! I hear you think A.I. is a bullshit generator. Well, we gave a whole profession of bullshit generators access to A.I., and you'll never believe how much more productive they became at generating bullshit! This is such a big deal for the Future of Work!" and goes on to encourage enthusiastic AI maximalists to think more carefully about the societal implications of these automations.
Which is precisely what Brian J. A. Boyd did in the beforementioned text. He not only observes a potential wave of automation for knowledge workers in all sectors but also notes a loss of the human sense of agency — the essential experience of human life itself. As humans, we can make plans, put them into action, and witness their effects. Replacing knowledge work with algorithms, automatizing the organizing, bureaucratic backbones of our society, might take away any sense of agency we have in such a highly organized and bureacratized society.
AI may not deplete humanity’s total stock of agency, but it could well redistribute it in an all-too-familiar fashion: Whoever has, will be given more; and whoever does not have, even that little shall be taken away. And life without agency is not merely joyless, it is an impoverished life. Prisoners in solitary confinement have been found to suffer physiologically as well as psychologically, and indeed, a basic cause of suicide is the belief that one is powerless to change one’s circumstances.
On the flip side, there is a joy or at least a justice in having what acts on us be an embodied agent. Much of the alienation we already experience in modern life comes from being pushed around without even the consolation of knowing that someone is doing the pushing. To put it in terms that corporations might understand: If the possibility of workers losing their purpose isn’t worrisome, the possibility of customers losing their minds should be. Even in a Kafkaesque encounter with bureaucracy, there is always a slim hope of appealing to the human being who still remains hidden within the job description, behind the desk. But when AI agents become capable of replacing white-collar jobs and acting with agency in social roles, that distinction between person and role is elided, and we go from René Magritte’s famous portrait The Son of Man, depicting the archetypal functionary as an anonymous man in a suit, face obscured by a floating apple, to the world of Norman Rockwell’s satirical Mr. Apple, which takes the same suited man but altogether replaces his head in favor of the fruit. We become confronted with a humanoid whose “face” itself is the API directions, the chat-box prompt, the disembodied AI voice in one’s ear.
Our lives in a modern, capitalist society have already significantly diminished this experience of agency. The rise of artificial intelligence, especially in administration, services and knowledge work, risks further eroding it.
Therefore, the liberation from seemingly burdensome and meaningless work is a double-edged sword when it condemns people to a life as an executive arm of a completely automated machine and its algorithmic instructions. Hence, we should be cautious about celebrating the end of so-called "bullshit jobs," as I have done with enthusiasm in the past because "uh i’m a creative and who needs bullshit jobs anyway", shrugging away the more dire consequences i have not thought through.
AI-Grift Shifts and the Content Creator Brain
A few weeks ago, Michelle Celarier wrote about what she calls the phenomenon of the "Grift Shift" in which venture capital and tech companies are redirecting their investments from the crypto hype and the now-fading metaverse into AI startups. A major beneficiary of this AI Grift Shift is Nvidia, which invests in AI-startups which in turn buy thousands of GPUs from them.
Celarier's prime example is a startup that, just a few years ago, rebranded itself from Applied Sciences to Applied Blockchain and now rebranded as Applied Digital providing AI-Hosting. This company is securing multi-million-dollar contracts with well-known AI brands like Character.ai and Stability, two of the major players in the generative AI business. They have ordered no less than 26,000 Nvidia GPUs, worth over a billion dollars.
Jürgen “tante“ Geuter, however, doesn't focus on this Grift Shift in the business side of tech, where companies behind the scenes shuffle backroom deals to keep the hype train going, and looks at a more fundamental problem with the seriousness of tech discourse, which has been infiltrated by shallow influencers. These influencers' primary goal is to create content and, similar to VC firms, jump every hype, shifting their grift.
Geuter defines this discursive Grift Shift as follows:
The Grift Shift is a new paradigm of debating technologies within a society that is based a lot less on the actual realistic use cases or properties of a certain technology but a surface level fascination with technologies but even more their narratives of future deliverance.
Within the Grift Shift paradigm the topics and technologies addressed are mere material for public personalities to continuously claim expertise and “thought leadership” in every cycle of the shift regardless of what specific technologies are being talked about.
According to Geuter, these influencers, largely shaped by YouTube culture, don't take their respective subjects seriously and mainly use them to feign non-existent expertise, whether it's about blockchains, productivity software, the metaverse, AI-art, large laguage models, or even room-temperature superconductors, depending on whatever is making the rounds. You can easily identify these grift shifters from their formulaic use of Twitter threads: “You’re missing out.“
Geuter observes that the tech discourse, which casually accepts or even elevates these guys as serious conversation partners, has reached an advanced stage of what he calls "content creator brain", which is the antithesis of an actual and serious engagement with the given topic. He’s right and I don’t think we’ve seen the climax of this development yet: As AI-tech now will be built into any and all consumer facing interfaces, this hypetrain might actually speed up and produce even more of those freeriding grifters than already clog the pipes.
A renewed seriousness in the general AI discourse on social media would quickly dispel debates about consciousness or creativity in statistical models and shift the focus toward discussions of the political and social impacts of these technologies, as well as their regulation. This more seriousness right now is pretty much exclusive to a more or less small circle of neo-ludites compared to the mass of shruggsters and hypesters: the ethicists, theoretical thinkers and tech activists — even when their voices get heard due to critical AI-reporting in the press. At the end of the day, one of the main jobs of journalism is a bone dry seriousness about it’s topics, isn’t it.
You wouldn’t think of it from a guy who uses headlines like Poltergeist Attacks on tripping Cars, but I'm always in favor of at least some seriousness in tech discourse. After all, the title of my very first selfproduced and handcoded digital magazine, which I invented at the tender age of 15 and published for an astounding two issues through the Amok release group on the C64, was: Serious by Damien/G*P.