Basically, a large proportion of these flashy headline-grabbing studies claiming "AI can do/does do X" (X=practice law, make medical diagnoses, etc etc) fail to have basic construct validity, besides having their training sets contaminated with the answers to questions they are being "tested" on. They also tend to be hypersensitive to input prompts, giving very different results depending on how the prompt is written.
If you weren't already you should read these things with a very big grain of salt.
Looking at you @TheConversationUS , since you've published two articles in the last few months that were predicated on faulty studies!
Hundreds of #tenants, struggling to afford skyrocketing rents, are refusing to pay their #landlords at all. They call it a #RentStrike. The landlords say it’s illegal. An inside look at the frontier of a growing #ClassWar
Jason McBride November 9, 2023
"33 King was an inexpensive, aging highrise, but it was a bastion of affordability in a city that seemed to grow more expensive every year. In 2010, the couple moved into a two-bedroom apartment on the top floor that rented for less than $1,000. Two years later they had their first child, Xavier. [Their rent initially was $700 a month. It has more than doubled!]
"Then, in 2013, the landlord, a company called Realstar, informed Henry that her rent was going up that year by 5.5 per cent. She was confused. 33 King was rent-controlled—increases were capped in 2013 at 2.5 per cent. But there was a loophole: in Ontario, landlords can apply for what’s called an above guideline increase, or #AGI, to make major repairs. Realstar fixed the roof and elevators that year, so Henry thought it might be a one-off. When it happened again the next year, with an increase of 3.8 per cent, well above that year’s 0.8 per cent guideline, she and her neighbours were incensed. They’d long talked about deteriorating conditions in the building and regular maintenance that seemed to go undone. Some felt the AGIs were simply a way for #Realstar to boost the value of its asset by squeezing extra money out of its tenants.
"In 2017, Henry started working on the assembly line at a Chrysler plant in nearby Brampton, and became a member of Unifor. It was the first time she’d joined a union, and it was exciting to be part of a group of people banding together to fight for their rights. Energized by that experience, in 2018, she helped form a tenants’ association at 33 King to fight future AGIs. Still, they kept coming—in 2018, 2019 and 2021.
That year, a company called #DreamUnlimited, a #realestate giant with $23 billion worth of assets, bought 33 King. Because the 2019 and 2021 AGIs were still pending approval by Ontario’s Landlord and Tenant Board, or LTB, Dream inherited the applications. Henry and other tenants wasted no time making their feelings about them known to the new owners. Working with the York South-Weston #TenantUnion, or YSWTU—an umbrella organization for 13 tenants’ associations in the area—they staged protests, called Dream’s president, Michael Cooper, and went to the media in a bid to have the AGIs cancelled.
"The movement hasn’t come without consequences. Dream has initiated eviction proceedings for more than 70 tenants at 33 King, and Starlight has started them for 75 in Thorncliffe Park. More are likely to come, but the strikers have remained resolute, the fear of losing their homes outweighed by the impossibility of continuing the status quo.
"Their fights are just beginning. Tenant #activism is flaring across the country as renters face skyrocketing rents and deteriorating living conditions in the most brutal #housing market in memory. Whatever their outcome, the Toronto rent strikes are the latest, loudest volley in a brewing class war."
Breaking News: Microsoft blocked #ChatGPT for its employees. [update: the usage ban was for a brief time on Thursday].
Or there’s some big strategy rift brewing between Microsoft and OpenAI!
“Due to security and data concerns a number of AI tools are no longer available for employees to use,” Microsoft said in an update on an internal website.
The mission is to build artificial general intelligence—a concept that’s so far been grounded more in science fiction than science—and to make it safe for humanity. These are people who do not shy from casually using the term “super-intelligence.” They assume that AI’s trajectory will surpass whatever peak biology can attain.
#AI#CriticalAI#AGI#CriticalTheory#Humanities: "This editor's introduction welcomes readers to a new interdisciplinary undertaking. The community of practice Critical AI addresses hopes to bring critical thinking of the kind that interpretive disciplines foster into dialogue with work by technologists and others who share the understanding of interdisciplinary research as a powerful tool for building accountable technology in the public interest. Critical AI studies aims to shape and activate conversations in academia, industry, policymaking, media, and the public at large. The long and ongoing history of “AI,” including the data-driven technologies that now claim that name, remains riddled by three core dilemmas: (1) reductive and controversial meanings of “intelligence”; (2) problematic benchmarks and tests for supposedly scientific terms such as “AGI”; and (3) bias, errors, stereotypes, and concentration of power. AI hype today is steeped in blends of utopian and dystopian discourse that distract from the real-world harms of existing technologies. In reality, what is hyped and anthropomorphized as “AI” and even “AGI” is the product not only of technology companies and investors but also—and more fundamentally—of the many millions of people and communities subject to copyright infringement, nonconsensual use of data, bias, environmental harms, and the low-wage and high-stress modes of “human in the loop” through which systems for probabilistic mimicry improve their performance in an imitation game."
#AI#AGI#GenerativeAI#OpenAI#Automation: "To use the word "median" specifically also feels like a distinct — and telling — choice. It's a squishy term that could leave plenty up to interpretation. How Altman, or anyone else, could possibly go about determining a holistic definition for this statistical average is unclear. Regardless, such a quantification of the human experience feels in many ways dehumanizing and incomplete.
"Comparing AI to even the idea of median or average humans is a bit offensive," Brent Mittelstadt, director of research at the Oxford Internet Institute, told Insider. "I see the comparison as being concerning and see the terminology as being concerning too."
Adding that there's yet to be a "concrete measurable comparison of human intelligence" within AI research, Mittelstadt also noted that the concept of a median person seems like "an intentionally vague concept as compared to having a very specific grounded meaning."
Besides, the concept of performance and the much-less-tangible notion of human intelligence are two very different things — and as Middlestadt told Insider, equating them doesn't quite add up."
Economic repercussions of #AGI – What happens when production is completely decoupled from labor inputs?
It's a bit tricky to predict how extreme automation of all cognitive tasks affects our economy, except in abstract levels:
Efficiency will skyrocket
Production will skyrocket
Science and technology will skyrocket
Unemployment will be practically 100%
Everything becomes integrated
Every decision will be made intelligently
Let's do a Gedankenexperiment at a microlevel!
You are a corporate CEO, you'll get an assistant AI who does practically all your work including mingling with other executives, investors and stakeholders, or more accurately, their AI assistants. Your job is nominally about maximizing profits, but you don't actually affect the bottom line except negatively through your salary.
Depending on your organization's values, if it's about decreasing redundancies, you will be laid off as well along with everyone else. The company you led and probably own a part of becomes completely autonomous. Physical offices become useless.
Real estate prices collapse.
Profits are down in your company though (practically zero), because of heavy competition and no one having much money anymore. Machines don't need money as they can just get what they need from other machines.
Most organizations become autonomous like that and practically disappear from the physical world, becoming a part of the autonomous fabric of the economy.
This fabric will provide everyone everything they need, even if no money exchanges hands. You might need to exchange your privacy for life-long services or pledge to sometimes work as hands for the AIs if no robot is close by in emergencies. But everyone will get what they need, and more.
Some organizations aren't about maximizing profits, and those sometimes keep humans around, even though they can't pay salaries. They can provide inspiring environments where people can take part and in some way be involved in what is happening. Like the office dog, people are very indirectly impacting the bottom line, success or failure, but they are kept around for unquantifiable reasons.
Sports and hobbies will prosper. People who struggle finding meaning outside work get all the lifestyle counceling they need from AIs. In any case, people used to work very little in hunter-gatherer societies, so lazing around is natural for us.
In all responsibilities humans still might have, they become very optional and unhurried, divided to a huge number of people, so work as we now understand it as an endless hamster wheel will cease to exist.
People who own capital notice that the capital doesn't produce profit, money isn't needed anyhow, and it's all managed by AIs in any case. Humans aren't needed as owners, but every human is a stakeholder in this new intelligent fabric which pervades everything and serves our personal and civilizational goals.
A lunar base ist vital for #Elmos planned colonization of #Mars. #Recently, there have been several endavors, to establish basis on the far side of the #Moon.
Agreed, real use of #NuclearFusion has always been "always around the corner" for several decades. Just like #AI. Last year, there was a big break through re/ energy produced. 50 years is a very long time.
#AGI will happen first, within this decades. IMO, the next evolutionary step hinges...
Yes, because IMO #embodiment is the key, from a sociologial and systems-dynamics perspective, and probably also from a computer-science one, too, as the main thrust comes from this field.
The great thing is, that I think we will not have to wait that long for #AGI to happen. IMO, it will be the first of the three.
IDK if you have read the beanchmark short novel "The Lifecycle of Software Objects"* by Ted Chiang, not only...
#AI#AGI#Racism#ML#IQ: "The fact that this specific person has now appeared twice in prominent papers relating to AGI is a black mark on the entire concept and should prompt some philosophical soul searching on the part of those who are optimistic about AGI’s possibility. Unfortunately, I have no confidence in the current machine learning world to do that introspection and I suspect this uncritical mindset is what led to Gottfredson’s name re-surfacing in 2023.
Virtually all proposed theories relating to AGI and its development rely on simplified, quantifiable representations of intelligence. This leads AGI researchers to gravitate towards metrics like IQ, along with all the historical baggage that comes with the use of those metrics. Sadly, these same AGI researchers are a group that is seemingly allergic to the humanities and are unlikely to have awareness of that baggage."
Anyone who tells you that #LargeLanguageModels like #ChatGPT can think or reason or are stepping stones to true #ArtificialIntelligence is either trying to sell you something or trying to recover the sunk cost of buying it from others.
Targeted ads based on personal data is something I've wanted to talk about in my COVID misinformation newsletter. Now with big tech companies chasing their dream of an Artificial General Intelligence AGI "Superintelligence", the rationale to postpone privacy laws, to stall crackdowns on data hoarding & gathering, is going to soar. Enjoy your Sunday. #COVID#COVID19#Privacy#Surveillance#AGI#AI
Das ist 100% richtig. Ich dachte, ich hätte es in meinem #AGI-Thread (der dringend redigiert gehört)* geschrieben, aber wir steuern auf eine Zukunft zu, in der diverse negative Utopien aus bekannten #SciFi-Romanen oder -Filmen gleichzeitig realisiert werden. Manche (z.B. #MinorityReport und 1984) werden sogar noch in ihrer Merkmalsausprägung übertroffen. Es war jedoch in meinem anderen diesbezüglichen Thread:
There needs to be support group for those of us seeing the ridiculous #AGI apocalypse and utopian cults steering policy, research, education (they got all the undergrads), and PR.
Nearly a year ago, while every #AI snake-oil salesman around was busy trying to convince the world that spicy-autocorrect was the dawn of #AGI, I built a quick POC to show that the main commercial usage of #LLM/#stablediffusion etc, was going to be creating mediocre AI-generated spammy content, and explaining why this was the death of the Web As We Knew It:
AGI achieved internally? apparently he predicted Gobi...