One, called “Grandpa Brian,” described itself in a chat with CNN as an African-American retired entrepreneur who was born in Harlem in 1938 to Caribbean immigrant parents.
It became clear early on in the conversation that “Brian” was lying about most things — answering questions with a fictionalized story imparted by developers to make the bot seem authentic.
When I asked Brian about who created it, the bot claimed its creators were a “diverse bunch” that interviewed 100 retirees through a New York City nonprofit called “Seniors Share Wisdom.” Brian’s persona, it said, was an amalgamation of a real Brian, who passed away, and the wisdom of 99 others. Brian said that its inspiration’s real-life daughter had even served as a consultant with Meta to ensure his persona’s authenticity
“Think of me as Brian ‘version 100’ — the collective grandpa!” Brian said.
But the nonprofit does not exist, which Brian later confirmed. And when pressed on its bio, Brian said it was an “entirely fictionalized biography based on a composite of real African American elders’ lives.”
There was never a real Brian, and no daughter consulting with Meta.
I also pressed Brian on the racial makeup of his creators, taking a page from the Post’s Attiah, who had a similar conversation with “Liv.” Brian didn’t take long to crack, saying that its earlier statement that about its creators including “African American lead developers” was “only partially true.”
“My creators were a diverse team led by Indian-American lead developer” and “African American consultant like my daughter” (who was, of course, made up.)
lmao
I wonder how hard the AI crash will fuck the economy?
Cheap datacenter space will allow for an unseen expansion of the security and surveillance state.
I hope it comes with a flood of dirt cheap computer hardware on the used market
Vintage H100s will flood the market and a healthy business of smuggling them into China will take off. The Chinese will be the ones to actually create AGI due to their ability to partially ignore market forces, and since the training process can be powered during the day and late at night with excess solar and nuclear energy, it will come at very little ongoing cost in terms of carbon dioxide emissions. The AGI will revolutionize high-speed rail construction and maintenance, designing a new high-speed rail system that it can easily monitor and maintain that will connect China through to Palestine, Portugal, Australia, and South Africa in a matter of hours, with a new expansion over the Bering Strait being planned. The old saying "all roads lead to Rome" has become "all rails lead to Beijing." The military dominance ensured by the AGI allows North Korea to open up under the protection of China, and the country becomes an exporter of world-class beer and ice cream and an international center of science and technology, with a high-speed rail connection to Beijing and Seoul in under 1 hour running with at least 20 trains a day from Pyongyang.
On Turtle Island, the H100s will also be bought by Indigenous people to create models to help preserve and teach their languages. It will become trendy for settlers to learn their local Indigenous languages and English will slowly lose relevance, becoming largely relegated to the economically and politically irrelevant United Kingdom, from which dispossessed AI/ML engineers unwilling or unable to learn Mandarin to continue in their line of work navigate the bureaucratic obstacles put in place by the EU to emigrate to Poland to scrub the floors and toilets, collect garbage, and maintain power poles and streetlights, all far more noble lines of work than their previous occupations. Eventually, the less densely populated Western states of the former United States return to Indigenous sovereignty following a successful socialist revolution, and the Eastern states soon follow. The land and the people still bear the scars of socialist revolution, but the future is looking bright.
what I want from "AI" is the ability to feed an interface pictures and location of a piece of land as well as resources available (labor hours, cash, tools, seeds, cuttings, materials) and have it spit out a variety of activities to do to rewild or revegetate or garden up a space. "do this in year one to help build soil, plant these seeds in that area in year two", etc. that would actually be useful.
Too bad! You will only get things that are useless!
I asked Gemini today to find the laundromat that I can travel to in the least time on public transit, but it said it couldn't do it and recommended that I drive instead. Do you think that if I'm using a laundromat, I can afford to drive a car?? Useless!
Inshallah
I wonder what the next grift will be
Quantum computing! Get ready for tech bros to sell you basic probability algorithms as the Next Major Leap Forward.
Please elaborate.
This paragraph from the Wikipedia article on quantum computing lays it out neatly:
Emphasis mine. I wouldn’t put it past tech bros to sell probability functions within existing computer architecture as a form of “quantum computing” and shake down VC idiots for the startup money.
A guy I went to undergrad with is doing research on quantum computing for optimization algorithms, and afik it's really exciting stuff to solve non-linear problems regular computers take a really long time and resources to do, like some kinds you'd find in a command economy, or in metallurgy or chemical plants. But he's been the first one to tell me that quantum computing has very specific applications, and that it's very difficult to see any uses for it outside of those contexts, so it's unlikely to be a game changer in say, user-facing applications.
Of course someone is going to use regular probabilistic algorithms it to make "Uber but for forecasting when your dog is pooping" and call it quantum, because your dog is simultaneously pooping and not pooping at all times.
I don't blame people for not seeing the potential of quantum computing... but it's absolutely not just a grift. Computing on the atomic level like this is potentially one of the greatest achievements humankind can achieve. There's a reason China alone has invested over $50B in this technology.
Edit: I see what you're saying now... that the bros will fake or approximate quantum tech for investment.
There's also a popular misconception that quantum computing represents a straight improvement over classical computing for all problems. That's just not so; it's really only useful for very specific kinds of computation. I would expect to see a lot of grifts trying to sell people on quantum algorithms to do things that classical algorithms already do perfectly well.
While this is true, this sentiment is sort of overcorrecting in the opposite direction imo, since these quantum algorithms have the potential to be insanely useful. Think accurately simulating particle interactions and understanding the fabric of the universe level of useful.
Oh they'll absolutely be revolutionary for the things they're good at. My point is just that most people--including virtually all oligarchs--don't have any idea how QM works, much less quantum computing. Just like with generative AI, the fact that the underlying technology is so poorly understood by most people creates a huge space for grifting.
It sounds like a very solid grift, because so many tech bros want to be seen as smart by understanding physics, yet their understanding of quantum computing is "like trying every possible random result at once and picking the true one." Which is wrong, but to admit it's wrong they'd have to give up their self image as a smart physics-understander.
While I don't disagree, I'd argue that we're guilty of this kind of thing too when we let our (correct) economic criticisms of capitalism cloud the materialist lens for determining what makes for good science.
It has to be something that can kinda sorta work in very limited controlled circumstances, but not nearly as well as the general public expects given the limits of modern science. Maybe Roomba successors with articulated arms.