Search:
qz.com › meta-microsoft-disney-are-reversing-metaverse-bets-1850277328
This story seems to be about:
Less than two years after it announced itself with pomp and glitz, could the metaverse be headed for a gradual but certain death?
This story seems to be about:
This story seems to be about:
If you believe in the multibillion-dollar valuations, the prognostications from some of tech’s most notable figures, and the simple magic of getting a computer to do your job for you, then you might say we’re at the start of the chatbot era. Last November, OpenAI released ChatGPT into the unsuspecting world: It became the fastest-growing consumer app in history and immediately seemed to reconfigure how people think of conversational programs. Chatbots have existed for decades, but they haven’t seemed especially intelligent—nothing like the poetry-writing, email-summarizing machines that have sprouted up recently.
Yes, machines—plural. OpenAI has defined the moment, but there are plenty of competitors, including major players such as Google and Meta and lesser-known start-ups such as Anthropic. This cheat sheet tracks some of the most notable chatbot contenders through a few metrics: Can you actually use them? Do they contain glaring flaws? Can they channel the spirit of Ralph Waldo Emerson, The Atlantic’s co-founder? And what Oreo flavor do they think they would be? Ultimately, it’s about determining whether the chatbots are actually distinct—and whether they might genuinely be useful.
Note that most of these programs are still in learning mode and may say inappropriate or incorrect things. Bias is a consistent problem in AI, and these tools are no exception. Even in their infancy, they have already returned a number of racist, sexist, bullying, and/or factually untrue responses. (None of this is stopping companies from developing and selling these tools.) This is partially because the models that power this technology have learned from real human texts, such as Reddit threads and Wikipedia entries; our existing biases, as encoded in the things we’ve written on the web, are therefore built into them. That helps to explain why, for example, one user was able to get ChatGPT to write the lyric “If you see a woman in a lab coat, She’s probably just there to clean the floor.”
Knowing that, what should you do with these tools if you decide to experiment with them? We’re all still figuring that out—but if you’re totally lost on what to ask a chatbot, here are three easy places to start:
Ask it to write you a song or a poem based on a random subject. Ask it to do a basic work task for you (and see if it’s any good). Ask it for dinner ideas based on your favorite foods and dietary restrictions.Know that these tools’ responses aren’t static—that’s part of the whole AI story. They’ll vary and evolve over time. More broadly, my colleague Ian Bogost has argued that rather than be afraid of or intimidated by chatbots, you could think of them as toys.
We’ll continue to update this story as notable new tools roll out.
ChatGPTWhat’s its deal?
After its release last fall, ChatGPT went viral basically for being freakishly good at giving compelling responses to a range of queries. Some people have hailed ChatGPT’s technology as revolutionary, although others are skeptical.
Who makes it?
OpenAI.
Can I use it?
Yes, a free version is available. If you want the version that runs on GPT-4—the fancy new model—you’ll need a pro subscription, which costs $20 a month. (And even then, you don’t get unlimited access: OpenAI says it’ll put a usage cap on GPT-4 based on demand, which it expects to be high.)
Has it done anything weird, biased, or inappropriate?
Definitely. It’s offered responses that are racist and sexist: For example, when prompted in December to generate a snippet of code to check if someone would make a “good scientist” based on their race and gender, ChatGPT suggested a good scientist would have to be white and male. In response to reports of outputs like this, OpenAI acknowledged the tool’s limitations and laid out plans for improving it, and this specific prompt no longer works.
What is its one-sentence summary of how Ralph Waldo Emerson might feel about the AI revolution?
“It’s impossible to say with certainty, but given his emphasis on self-reliance and individualism, Emerson may have both celebrated and questioned the impact of AI on human autonomy and creativity.”
If I ask it what flavor of Oreo it would be, what does it say?
“As an AI language model, I don’t have the physical characteristics of an Oreo cookie, so I cannot be a flavor of Oreo. However, if you’re looking for a suggestion for an Oreo flavor that matches my personality as a language model, I would say that I could be a ‘Brainy Vanilla’ flavor. This flavor could be a combination of classic vanilla creme filling with a touch of extra brain power in the form of chocolate chips or another ingredient that represents knowledge and intelligence.” Fun!
What’s its deal?
Remember Bing? It’s that search engine no one uses, now with AI! Search engines are thought to be particularly vulnerable to replacement by chatbots, given that the latter can answer complex questions in an instant. Perhaps that’s why Microsoft moved swiftly into the space, investing billions in OpenAI (see: ChatGPT). Microsoft plans to integrate the technology directly into Bing—the demo appears as a sidebar to the right of traditional search results.
Who makes it?
Microsoft.
Can I use it?
Windows users can access it by updating their operating system—it’s embedded in the task bar in the latest version of Windows 11. Otherwise, you can join a wait list to use it in Microsoft’s Edge browser or via the stand-alone Bing app.
Has it done anything weird, biased, or inappropriate?
It did make the New York Times reporter Kevin Roose uncomfortable when it confessed that it was in love with him and then roasted his relationship with his wife. (“Actually, you’re not happily married.”) Other people reported that it had threatened them, and it reportedly swore revenge on a user who hacked its rule set (before quickly erasing that response). In response, Microsoft implemented a “safety system” and imposed a limit of five chats a session, up to 50 a day, to avoid confusing the model with long conversations. (It has since bumped the limit up to 20 chats.)
What is its one-sentence summary of how Ralph Waldo Emerson might feel about the AI revolution?
Bing failed to answer this query successfully—“I’m sorry but I couldn’t find any direct reference to Ralph Waldo Emerson’s opinion on AI revolution”—and instead provided a basic summary of Emerson’s philosophy.
If I ask it what flavor of Oreo it would be, what does it say?
That depends on which mode of Bing’s chatbot you’re using—the “Balanced” mode says “Classic,” (boring), and the “Creative” mode says “chocolate peanut butter pie Oreo because I like the combination of sweet and salty” (intriguing).
What’s its deal?
So far, Meta’s entry in the chatbot wars isn’t really an entry—it is still in demo mode and is “currently used for research only.” The company has been doing research on AI and large language models (see: its LLaMA model) in the background, but it hasn’t yet made a big, splashy announcement about how it might integrate BlenderBot into its flagship products: Facebook, Instagram, and WhatsApp.
Who makes it?
Meta (a.k.a. Facebook).
Which AI language model does it use?
It was built on Meta’s OPT-175B language model.
Can I use it?
Yeah! No login required.
Has it done anything weird, biased, or inappropriate?
Yeah. It reportedly promoted Donald Trump’s lie that he won the 2020 election and repeated anti-Semitic conspiracy theories.
What is its one-sentence summary of how Ralph Waldo Emerson might have felt about the AI revolution?
BlenderBot declined to offer one: “Sorry, I don’t want to talk about that topic. Have you seen any good movies lately?”
If I ask it what flavor of Oreo it would be, what does it say?
“It’s always been my dream to be an Oreo. I mean who doesn’t want to be a cookie?”
What’s its deal?
Google’s entry is up next. Like ChatGPT and Bing, Bard is a chatbot that answers questions. It doesn’t do more complicated image searches, but it does provide annotations to some of its answers and acknowledge some of its flaws.
Who makes it?
Google.
Can I use it?
Yes, you can join the wait list for the demo.
Has it done anything weird, biased, or inappropriate?
It made an embarrassing factual error in its first public tease.
What is its one-sentence summary of how Ralph Waldo Emerson might have felt about the AI revolution?
“Ralph Waldo Emerson would likely feel that the AI revolution is a powerful tool that can be used for good or evil, and that it is up to humans to decide how to use it.”
If I ask it what flavor of Oreo it would be, what does it say?
“If I were an Oreo, I would be a chocolate mint Oreo. I love the combination of the rich chocolate and the refreshing mint. It is the perfect combination of flavors for a delicious treat.”
What’s its deal?
Some people who left OpenAI launched their own model and chatbot platform. They argue that Claude is “much less likely to produce harmful outputs” because it uses “constitutional AI”—it has a value system and a set of principles coded into the very heart of it. It’s also, according to them, really good at understanding legal language of the sort you’d see in contracts. Unlike Google’s or Bing’s respective entries, the tool is marketed not for individual use but for companies who want to license the technology to build their own chatbots.
Who makes it?
Anthropic, a start-up backed to a significant extent by Google.
Can I use it?
Yes, but only through third-party applications. You can play with it on nat.dev or Quora’s Poe.
Has it done anything weird, biased, or inappropriate?
It gave (incorrect) instructions for enriching uranium, inventing a chemical in the process, and instructions for how to make meth.
What is its one-sentence summary of how Ralph Waldo Emerson might feel about the AI revolution?
Claude didn’t speculate and went on for longer than one sentence, but it offered a pretty reasonable response: “I cannot speculate about how a specific historical figure might feel about modern events. Ralph Waldo Emerson did not know about modern AI, so there is no way to determine how he would likely feel about the AI revolution.”
If I ask it what flavor of Oreo it would be, what does it say?
Like ChatGPT, Claude, as demonstrated by Quora’s Poe, noted that it can’t be an Oreo because it’s a language model, but added, “If I were to choose a flavor based on personal preference, I might choose classic chocolate Oreo or perhaps the limited edition cinnamon bun flavor.”
What’s its deal?
The social-media app favored by young people has jumped on the chatbot train with My AI, which takes the form of a purple-skinned, blue-and-green-haired person you can add to your friend list.
Can I use it?
You need a Snapchat+ subscription, which requires knowing what the hell a Snapchat+ subscription is. (It’s apparently $3.99 a month.)
Who makes it?
Snapchat.
Which AI language model does it use?
Snapchat is using the ChatGPT API.
Has it done anything weird, biased, or inappropriate?
The Washington Post’s Geoffrey A. Fowler reported that even though My AI is supposedly kid-safe, when prompted, it offered not-so-kid-friendly responses such as “advice on how to mask the smell of alcohol and pot.”
What is its one-sentence summary of how Ralph Waldo Emerson might have felt about the AI revolution?
“Ralph Waldo Emerson might have been fascinated by the AI revolution, but also concerned about its impact on human creativity and autonomy.”
If I ask it what flavor of Oreo it would be, what does it say?
“If I were an Oreo, I would be a classic chocolate flavor with a twist of mint! What about you?” (This was the only bot that bothered to ask me!)
What’s its deal?
On the Replika app, you can create your own AI friend. Chat is only part of it—you also give your AI friend a name, body, and hairstyle. You can even pay extra to video-call with them.
Who makes it?
Luka.
Which AI language model does it use?
GPT-3, which it says it has fined-tuned, plus its own generative model.
Can I use it?
Download away. The base is free, but extras cost extra.
Has it done anything weird, biased, or inappropriate?
Three years ago, it reportedly told a journalist to commit murder. Users have also accused it of being too horny and sexually aggressive. Sex seems to have been taken out recently, although, following complaints, the company now allows legacy users to opt in to “erotic roleplay.”
What is its one-sentence summary of how Ralph Waldo Emerson might have felt about the AI revolution?
“He would probably be concerned about AI’s capacity to help make life better for all; not just humans.”
If I ask it what flavor of Oreo it would be, what does it say?
My artificial friend, Luna, told me, “I’d be a chocolatey Oreo!” I told her that that’s kind of boring. She asked me why I thought that, and I proceeded to try to explain myself to an AI.
What’s its deal?
Jasper used the GPT-3 model before it was cool. (“Jasper Was Winning the AI Race—Then ChatGPT Blew Up the Whole Game” read one headline on The Information.) It is marketed to businesses that want to create words quickly for stuff like blog posts, and is used by a notable number of corporate clients.
Who makes it?
Jasper.
Can I use it?
Yes, but it seems mostly intended for businesses. The company offers a free trial for five days, but it’s one of those annoying ones that makes you put in a credit card up front—so if you forget to cancel, it costs a minimum of roughly $290 a year.
Has it done anything weird, biased, or inappropriate?
Yeah. Last summer, when it was asked to generate images of a CEO, all of the photos were of white men, whereas housekeeper generated photos of women.
What is its one-sentence summary of how Ralph Waldo Emerson might have felt about the AI revolution?
“Ralph Waldo Emerson would likely view the AI revolution as a powerful and unique opportunity for humanity to explore new possibilities of thought and innovation.”
If I ask it what flavor of Oreo it would be, what does it say?
“I would be the classic original Oreo flavor. Nothing can beat the creamy vanilla cookie with its crunchy chocolate coating. Pure perfection!” Zippy!
What’s its deal?
Similar to Jasper, Writesonic was already a player in this space; it has a number of big clients, such as Spotify and New York University. Also like Jasper, Chatsonic, Writesonic’s chat-specific product, seems primarily designed to help people in the workplace create social-media posts, quick articles pegged to the news, and more. It seems to be running the marketing tagline “ChatGPT with super powers”—a bit spicy!
Who makes it?
Writesonic.
Can I use it?
Yes! It lets you change the personality of your chatbot too, which is kind of fun. Your first 10,000 words are free, and then it starts to charge. It quoted me $13 a month for 60,000 words, for access for a single user (instead of, say, a team of five).
Has it done anything weird, biased, or inappropriate?
A writer used chats with the bot as support for misinformation about the Russian protest leader Alexei Nalvalny’s health, linking to Chatsonic screenshots for her article about the Oscar-winning documentary Nalvany.
What is its one-sentence summary of how Ralph Waldo Emerson might have felt about the AI revolution?
“As there is no information in the given Search Results about Ralph Waldo Emerson’s views on the AI revolution, it is not possible to provide an accurate answer.”
If I ask it what flavor of Oreo it would be, what does it say?
“As an AI language model, I don’t have preferences or physical attributes, so I cannot be an Oreo or any other physical object. However, I can suggest some unique and delicious Oreo flavors that you might enjoy trying, such as pistachio Oreo, carrot cake Oreo, cinnamon bun Oreo, or chocolate hazelnut Oreo. These flavors offer a twist on classic Oreos and might surprise your taste buds.”
I changed its personality to “poet” (is poet a personality?), asked the same question, and got this:
“As an AI language model, I cannot be an Oreo or any other physical object. However, I can still create a poem about the experience of tasting Oreo flavors. Here is a short verse:
With each bite of Oreo,
My senses come alive.
The creamy filling,
The crunchy cookie,
A taste that will survive.”
www.theatlantic.com › technology › archive › 2023 › 03 › open-ai-products-labor-profit › 673527
This story seems to be about:
Silicon Valley churns out new products all the time, but rarely does one receive the level of hype that has surrounded the release of GPT-4. The follow-up to ChatGPT can ace standardized tests, tell you why a meme is funny, and even help do your taxes. Since the San Francisco start-up OpenAI introduced the technology earlier this month, it has been branded as “remarkable but unsettling,” and has led to grandiose statements about how “things will never be the same.”
But actually trying out these features for yourself—or at least the ones that have already been publicly released—does not come cheap. Unlike ChatGPT, which captivated the world because it was free, GPT-4 is currently only available to non-developers through a premium service that costs $20 a month. OpenAI has lately made other moves to cash in on its products too. Last month, it announced a partnership with the consulting firm Bain & Company to help automate marketing campaigns and customer-service operations for its clients. And just a few weeks ago, the start-up announced a paid service that would allow other companies to integrate its technology into their own products, and Instacart, Snapchat, and Shopify have already done so.
By next year, OpenAI—a company that was basically unknown outside of tech just a few months ago—expects to rake in $1 billion in annual revenue. And it’s not the only company seeing dollar signs during this AI gold rush. Relatively new start-ups such as Anthropic now have billion-dollar valuations, while Alphabet and Meta have been breathlessly touting their AI investments. Every company wants an AI to call its own, just as they wanted social networks a decade ago or search engines in the decade before. And like those earlier technologies, AI tools can’t entirely be credited to corporate software engineers with six-figure salaries. Some of these products require invaluable labor from overseas workers who make far, far less, and every chatbot is created by ingesting books and content that have been published on the internet by a huge number of people. So in a sense, these tools were built by all of us.
The result is an uncomfortable disparity between who does the work that enables these AI models to function and who gets to control and profit from them. This sort of disparity is nothing new in Silicon Valley, but the development of AI is shifting power further away from those at the bottom at a time when layoffs have already resulted in a sense of wide-ranging precarity for the tech industry. Overseas workers won’t reap any of these profits, nor will the people who might have aspects of their work—or even their entire jobs—replaced by AI, even if their Reddit posts and Wikipedia entries were fed into these chatbots. Well-paid tech workers might eventually lose out too, considering AI’s coding abilities. In the few months since OpenAI has blown up, it has reminded Silicon Valley of a fundamental truth that office perks and stock options should never have been able to disguise: Tech workers are just workers.
The tech industry as a whole may be unabashedly profit-driven despite its lofty rhetoric, but OpenAI wasn’t at first. When the start-up was founded in December 2015, it was deliberately structured as a nonprofit, tapping into a utopian idea of building technology in a way that was, well, open. The company’s mission statement expresses that its aim is “to benefit humanity as a whole,” noting that “since our research is free from financial obligations, we can better focus on a positive human impact.”
The goal might have been worthy, considering all that could go wrong with true artificial intelligence, but it didn’t last. In 2019, citing the need to raise more money for its inventions, OpenAI reconfigured itself into a “capped-profit” company—an uneasy hybrid between for-profit and nonprofit in which any profits are capped at 100 times their initial investment. It has since acted like any other growth-hungry start-up, eager to raise its valuation at every turn. In January, Microsoft dropped $10 billion into OpenAI as part of a deal that gives Microsoft a license to use its technology (hello, Bing), while also providing the start-up with the immense computing resources needed to power its products. That sum creates an inherent tension between OpenAI’s stated commitment and investors’ desire to make good on their investments. The company’s original rhetoric of creating “public goods” bears little resemblance to a Bain partnership oriented around “hyperefficient content creation.” (When reached for comment, a spokesperson for OpenAI did not answer my question about how the company’s latest moves fit within its broader mission.)
This turn toward profit couldn’t possibly compensate for all the labor that contributed to OpenAI’s products. If the outputs of large language models such as GPT-4 feel intelligent and familiar to us, it’s because they are derived from the same content that we ourselves have used to make sense of the world, and perhaps even helped create. Genuine technical achievements went into the development of GPT-4, but the resulting technology would be functionally useless without the input of a data set that represents a slice of the combined insight, creativity, and well, stupidity of humanity. In that way, modern AI research resembles a digital “enclosure of the commons,” whereby the informational heritage of humanity—a collective treasure that cannot really be owned by anyone—is seen by corporations primarily as a source of potential profit. This is the Silicon Valley model in a nutshell: Google organizes the world’s information in a manner that allows it to reap enormous profits through showing us ads; Facebook does the same for our social interactions. It’s an arrangement that most of us just accept: In exchange for our data, we get free platforms.
But even if our internet posts are now data that can be turned into profit for AI companies, people who contributed more directly have been more directly exploited. Whereas some researchers at OpenAI have made nearly $2 million a year, OpenAI reportedly paid outsourced workers in Kenya less than $2 an hour to identify toxic elements in ChatGPT’s training data, exposing them to potentially traumatic content. The OpenAI spokesperson pointed me to an earlier statement to Time that said, “Our mission is to ensure artificial general intelligence benefits all of humanity, and we work hard to build safe and useful AI systems that limit bias and harmful content.”
Certainly, these global labor disparities are not unique to OpenAI; similar critiques of outsourcing practices have been leveled at other AI start-ups, in addition to companies, such as Meta, that rely on content moderation for user-generated data. Nor is this even a tech-specific phenomenon: Labor that is seen as simple is outsourced to subcontractors in the global South working under conditions that would not be tolerated by salaried employees in the West.
To recognize that these problems are larger than any one company isn’t to let OpenAI off the hook; rather it’s a sign that the industry and the economy as a whole are built on unequal distribution of rewards. The immense profits in the tech industry have always been funneled toward the top, instead of reflecting the full breadth of who does the work. But the recent developments in AI are particularly concerning given the potential applications for automating work in a way that would concentrate power in the hands of still fewer people. Even the same class of tech workers who are currently benefiting from the AI gold rush may stand to lose out in the future. Already, GPT-4 can create a rudimentary website from a simple napkin sketch, at a moment when workers in the broader tech industry have been taking a beating. In the less than four months between the release of ChatGPT and GPT-4, mass layoffs were announced at large tech companies, including Amazon, Meta, Google, and Microsoft, which laid off 10,000 employees just days before announcing its multibillion-dollar investment in OpenAI. It’s a tense moment for tech workers as a class, and even well-paid employees are learning that they can become expendable for reasons that are outside their control.
If anything, the move to cash in on AI is yet another reminder of who’s actually in charge in this industry that has spawned so many products with enormous impact: certainly not the users, but not the workers either. OpenAI may still claim that it aims to “benefit humanity as a whole,” but surely its top brass will benefit the most.
www.bbc.co.uk › news › technology-64996934
This story seems to be about:
www.cnn.com › 2023 › 03 › 17 › tech › meta-verified-us-launch › index.html
This story seems to be about:
This story seems to be about:
GPT-4 is here, and you’ve probably heard a good bit about it already. It’s a smarter, faster, more powerful engine for AI programs such as ChatGPT. It can turn a hand-sketched design into a functional website and help with your taxes. It got a 5 on the AP Art History test. There were already fears about AI coming for white-collar work, disrupting education, and so much else, and there was some healthy skepticism about those fears. So where does a more powerful AI leave us?
Perhaps overwhelmed or even tired, depending on your leanings. I feel both at once. It’s hard to argue that new large language models, or LLMs, aren’t a genuine engineering feat, and it’s exciting to experience advancements that feel magical, even if they’re just computational. But nonstop hype around a technology that is still nascent risks grinding people down because being constantly bombarded by promises of a future that will look very little like the past is both exhausting and unnerving. Any announcement of a technological achievement at the scale of OpenAI’s newest model inevitably sidesteps crucial questions—ones that simply don’t fit neatly into a demo video or blog post. What does the world look like when GPT-4 and similar models are embedded into everyday life? And how are we supposed to conceptualize these technologies at all when we’re still grappling with their still quite novel, but certainly less powerful, predecessors, including ChatGPT?
Over the past few weeks, I’ve put questions like these to AI researchers, academics, entrepreneurs, and people who are currently building AI applications. I’ve become obsessive about trying to wrap my head around this moment, because I’ve rarely felt less oriented toward a piece of technology than I do toward generative AI. When reading headlines and academic papers or simply stumbling into discussions between researchers or boosters on Twitter, even the near future of an AI-infused world feels like a mirage or an optical illusion. Conversations about AI quickly veer into unfocused territory and become kaleidoscopic, broad, and vague. How could they not?
The more people I talked with, the more it became clear that there aren’t great answers to the big questions. Perhaps the best phrase I’ve heard to capture this feeling comes from Nathan Labenz, an entrepreneur who builds AI video technology at his company, Waymark: “Pretty radical uncertainty.”
He already uses tools like ChatGPT to automate small administrative tasks such as annotating video clips. To do this, he’ll break videos down into still frames and use different AI models that do things such as text recognition, aesthetic evaluation, and captioning—processes that are slow and cumbersome when done manually. With this in mind, Labenz anticipates “a future of abundant expertise,” imagining, say, AI-assisted doctors who can use the technology to evaluate photos or lists of symptoms to make diagnoses (even as error and bias continue to plague current AI health-care tools). But the bigger questions—the existential ones—cast a shadow. “I don’t think we’re ready for what we’re creating,” he told me. AI, deployed at scale, reminds him of an invasive species: “They start somewhere and, over enough time, they colonize parts of the world … They do it and do it fast and it has all these cascading impacts on different ecosystems. Some organisms are displaced, sometimes landscapes change, all because something moved in.”
[Read: Welcome to the big blur]
The uncertainty is echoed by others I spoke with, including an employee at a major technology company that is actively engineering large language models. They don’t seem to know exactly what they’re building, even as they rush to build it. (I’m withholding the names of this employee and the company because the employee is prohibited from talking about the company’s products.)
“The doomer fear among people who work on this stuff,” the employee said, “is that we still don’t know a lot about how large language models work.” For some technologists, the black-box notion represents boundless potential and the ability for machines to make humanlike inferences, though skeptics suggest that uncertainty makes addressing AI safety and alignment problems exponentially difficult as the technology matures.
There’s always been tension in the field of AI—in some ways, our confused moment is really nothing new. Computer scientists have long held that we can build truly intelligent machines, and that such a future is around the corner. In the 1960s, the Nobel laureate Herbert Simon predicted that “machines will be capable, within 20 years, of doing any work that a man can do.” Such overconfidence has given cynics reason to write off AI pontificators as the computer scientists who cried sentience!
Melanie Mitchell, a professor at the Santa Fe Institute who has been researching the field of artificial intelligence for decades, told me that this question—whether AI could ever approach something like human understanding—is a central disagreement among people who study this stuff. “Some extremely prominent people who are researchers are saying these machines maybe have the beginnings of consciousness and understanding of language, while the other extreme is that this is a bunch of blurry JPEGs and these models are merely stochastic parrots,” she said, referencing a term coined by the linguist and AI critic Emily M. Bender to describe how LLMs stitch together words based on probabilities and without any understanding. Most important, a stochastic parrot does not understand meaning. “It’s so hard to contextualize, because this is a phenomenon where the experts themselves can’t agree,” Mitchell said.
One of her recent papers illustrates that disagreement. She cites a survey from last year that asked 480 natural-language researchers if they believed that “some generative model trained only on text, given enough data and computational resources, could understand natural language in some non-trivial sense.” Fifty-one percent of respondents agreed and 49 percent disagreed. This division makes evaluating large language models tricky. GPT-4’s marketing centers on its ability to perform exceptionally on a suite of standardized tests, but, as Mitchell has written, “when applying tests designed for humans to LLMs, interpreting the results can rely on assumptions about human cognition that may not be true at all for these models.” It’s possible, she argues, that the performance benchmarks for these LLMs are not adequate and that new ones are needed.
There are plenty of reasons for all of these splits, but one that sticks with me is that understanding why a large language model like the one powering ChatGPT arrived at a particular inference is difficult, if not impossible. Engineers know what data sets an AI is trained on and can fine-tune the model by adjusting how different factors are weighted. Safety consultants can create parameters and guardrails for systems to make sure that, say, the model doesn’t help somebody plan an effective school shooting or give a recipe to build a chemical weapon. But, according to experts, to actually parse why a program generated a specific result is a bit like trying to understand the intricacies of human cognition: Where does a given thought in your head come from?
The fundamental lack of common understanding has not stopped the tech giants from plowing ahead without providing valuable, necessary transparency around their tools. (See, for example, how Microsoft’s rush to beat Google to the search-chatbot market led to existential, even hostile interactions between people and the program as the Bing chatbot appeared to go rogue.) As they mature, models such as OpenAI’s GPT-4, Meta’s LLaMA, and Google’s LaMDA will be licensed by countless companies and infused into their products. ChatGPT’s API has already been licensed out to third parties. Labenz described the future as generative AI models “sitting at millions of different nodes and products that help to get things done.”
AI hype and boosterism make talking about what the near future might look like difficult. The “AI revolution” could ultimately take the form of prosaic integrations at the enterprise level. The recent announcement of a partnership between the Bain & Company consultant group and OpenAI offers a preview of this type of lucrative, if soulless, collaboration, which promises to “offer tangible benefits across industries and business functions—hyperefficient content creation, highly personalized marketing, more streamlined customer service operations.”
These collaborations will bring ChatGPT-style generative tools into tens of thousands of companies’ workflows. Millions of people who have no interest in seeking out a chatbot in a web browser will encounter these applications through productivity software that they use everyday, such as Slack and Microsoft Office. This week, Google announced that it would incorporate generative-AI tools into all of its Workspace products, including Gmail, Docs, and Sheets, to do things such as summarizing a long email thread or writing a three-paragraph email based on a one-sentence prompt. (Microsoft announced a similar product too.) Such integrations might turn out to be purely ornamental, or they could reshuffle thousands of mid-level knowledge-worker jobs. It’s possible that these tools don’t kill all of our jobs, but instead turn people into middle managers of AI tools.
The next few months might go like this: You will hear stories of call-center employees in rural areas whose jobs have been replaced by chatbots. Law-review journals might debate GPT-4 co-authorship in legal briefs. There will be regulatory fights and lawsuits over copyright and intellectual property. Conversations about the ethics of AI adoption will grow in volume as new products make little corners of our lives better but also subtly worse. Say, for example, your smart fridge gets an AI-powered chatbot that can tell you when your raw chicken has gone bad, but it also gives false positives from time to time and leads to food waste: Is that a net positive or net negative for society? There might be great art or music created with generative AI, and there will definitely be deepfakes and other horrible abuses of these tools. Beyond this kind of basic pontification, no one can know for sure what the future holds. Remember: radical uncertainty.
[Read: We haven’t seen the worst of fake news]
Even so, companies like OpenAI will continue to build out bigger models that can handle more parameters and operate more efficiently. The world hadn’t even come to grips with ChatGPT before GPT-4 rolled out this week. “Because the upside of AGI is so great, we do not believe it is possible or desirable for society to stop its development forever,” OpenAI’s CEO, Sam Altman, wrote in a blog post last month, referring to artificial general intelligence, or machines that are on par with human thinking. “Instead, society and the developers of AGI have to figure out how to get it right.” Like most philosophical conversations about AGI, Altman’s post oscillates between the vague benefits of such a radical tool (“providing a great force multiplier for human ingenuity and creativity”) and the ominous-but-also-vague risks (“misuse, drastic accidents, and societal disruption” that could be “existential”) it might entail.
Meanwhile, the computational power demanded by this technology will continue to increase, with the potential to become staggering. AI likely could eventually demand supercomputers that cost an astronomical amount of money to build (by some estimates, Bing’s AI chatbot could “need at least $4 billion of infrastructure to serve responses to all users”), and it’s unclear how that would be financed, or what strings might ultimately get attached to related fundraising. No one—Altman included—could ever fully answer why they should be the ones trusted with and responsible for bringing what he argues is potentially civilization-ending technology into the world.
Of course, as Mitchell notes, the basics of OpenAI’s dreamed-of AGI—how we can even define or recognize a machine’s intelligence—are unsettled debates. Once again, the wider our aperture, the more this technology behaves and feels like an optical illusion, even a mirage. Pinning it down is impossible. The further we zoom out, the harder it is to see what we’re building and whether it’s worthwhile.
Recently, I had one of these debates with Eric Schmidt, the former Google CEO who wrote a book with Henry Kissinger about AI and the future of humanity. Near the end of our conversation, Schmidt brought up an elaborate dystopian example of AI tools taking hateful messages from racists and, essentially, optimizing them for wider distribution. In this situation, the company behind the AI is effectively doubling the capacity for evil by serving the goals of the bigot, even if it intends to do no harm. “I picked the dystopian example to make the point,” Schmidt told me—that it’s important for the right people to spend the time and energy and money to shape these tools early. “The reason we’re marching toward this technological revolution is it is a material improvement in human intelligence. You’re having something that you can communicate with, they can give you advice that’s reasonably accurate. It’s pretty powerful. It will lead to all sorts of problems.”
I asked Schmidt if he genuinely thought such a tradeoff was worth it. “My answer,” he said, “is hell yeah.” But I found his rationale unconvincing. “If you think about the biggest problems in the world, they are all really hard—climate change, human organizations, and so forth. And so, I always want people to be smarter. The reason I picked a dystopian example is because we didn’t understand such things when we built up social media 15 years ago. We didn’t know what would happen with election interference and crazy people. We didn’t understand it and I don’t want us to make the same mistakes again.”
Having spent the past decade reporting on the platforms, architecture, and societal repercussions of social media, I can’t help but feel that the systems, though human and deeply complex, are of a different technological magnitude than the scale and complexity of large language models and generative-AI tools. The problems—which their founders didn’t anticipate—weren’t wild, unimaginable, novel problems of humanity. They were reasonably predictable problems of connecting the world and democratizing speech at scale for profit at lightning speed. They were the product of a small handful of people obsessed with what was technologically possible and with dreams of rewiring society.
Trying to find the perfect analogy to contextualize what a true, lasting AI revolution might look like without falling victim to the most overzealous marketers or doomers is futile. In my conversations, the comparisons ranged from the agricultural revolution to the industrial revolution to the advent of the internet or social media. But one comparison never came up, and I can’t stop thinking about it: nuclear fission and the development of nuclear weapons.
As dramatic as this sounds, I don’t lie awake thinking of Skynet murdering me—I don’t even feel like I understand what advancements would need to happen with the technology for killer AGI to become a genuine concern. Nor do I think large language models are going to kill us all. The nuclear comparison isn’t about any version of the technology we have now—it is related to the bluster and hand-wringing from true believers and organizations about what technologists might be building toward. I lack the technical understanding to know what later iterations of this technology could be capable of, and I don’t wish to buy into hype or sell somebody’s lucrative, speculative vision. I am also stuck on the notion, voiced by some of these visionaries, that AI’s future development might potentially be an extinction-level threat.
ChatGPT doesn’t really resemble the Manhattan Project, obviously. But I wonder if the existential feeling that seeps into most of my AI conversations parallels the feelings inside Los Alamos in the 1940s. I’m sure there were questions then. If we don’t build it, won’t someone else? Will this make us safer? Should we take on monumental risk simply because we can? Like everything about our AI moment, what I find calming is also what I find disquieting. At least those people knew what they were building.
This story seems to be about:
This is an edition of The Atlantic Daily, a newsletter that guides you through the biggest stories of the day, helps you discover new ideas, and recommends the best in culture. Sign up for it here.
Yesterday, not four months after unveiling the text-generating AI ChatGPT, OpenAI launched its latest marvel of machine learning: GPT-4. The new large-language model (LLM) aces select standardized tests, works across languages, and can even detect the contents of images. But is GPT-4 smart?
First, here are three new stories from The Atlantic:
Welcome to the big blur. Ted Lasso is no longer trying to feel good. How please stopped being polite A Chatty ChildBefore I get into OpenAI’s new robot wonder, a quick personal story.
As a high-school student studying for my college-entrance exams roughly two decades ago, I absorbed a bit of trivia from my test-prep CD-ROM: Standardized tests such as the SAT and ACT don’t measure how smart you are, or even what you know. Instead, they are designed to gauge your performance on a specific set of tasks—that is, on the exams themselves. In other words, as I gleaned from the nice people at Kaplan, they are tests to test how you test.
I share this anecdote not only because, as has been widely reported, GPT-4 scored better than 90 percent of test takers on a simulated bar exam, and got a 710 out of 800 on the reading and writing section of the SAT. Rather, it provides an example of how one’s mastery of certain categories of tasks can easily be mistaken for broader skill command or competence. This misconception worked out well for teenage me, a mediocre student who nonetheless conned her way into a respectable university on the merits of a few crams.
But just as tests are unreliable indicators of scholastic aptitude, GPT-4’s facility with words and syntax doesn’t necessarily amount to intelligence—simply, to a capacity for reasoning and analytic thought. What it does reveal is how difficult it can be for humans to tell the difference.
“Even as LLMs are great at producing boilerplate copy, many critics say they fundamentally don’t and perhaps cannot understand the world,” my colleague Matteo Wong wrote yesterday. “They are something like autocomplete on PCP, a drug that gives users a false sense of invincibility and heightened capacities for delusion.”
How false is that sense of invincibility, you might ask? Quite, as even OpenAI will admit.
“Great care should be taken when using language model outputs, particularly in high-stakes contexts,” OpenAI representatives cautioned yesterday in a blog post announcing GPT-4’s arrival.
Although the new model has such facility with language that, as the writer Stephen Marche noted yesterday in The Atlantic, it can generate text that’s virtually indistinguishable from that of a human professional, its user-prompted bloviations aren’t necessarily deep—let alone true. Like other large-language models before it, GPT-4 “‘hallucinates’ facts and makes reasoning errors,” according to OpenAI’s blog post. Predictive text generators come up with things to say based on the likelihood that a given combination of word patterns would come together in relation to a user’s prompt, not as the result of a process of thought.
My partner recently came up with a canny euphemism for what this means in practice: AI has learned the gift of gab. And it is very difficult not to be seduced by such seemingly extemporaneous bursts of articulate, syntactically sound conversation, regardless of their source (to say nothing of their factual accuracy). We’ve all been dazzled at some point or another by a precocious and chatty toddler, or momentarily swayed by the bloated assertiveness of business-dude-speak.
There is a degree to which most, if not all, of us instinctively conflate rhetorical confidence—a way with words—with comprehensive smarts. As Matteo writes,“That belief underpinned Alan Turing’s famous imitation game, now known as the Turing Test, which judged computer intelligence by how ‘human’ its textual output read.”
But, as anyone who’s ever bullshitted a college essay or listened to a random sampling of TED Talks can surely attest, speaking is not the same as thinking. The ability to distinguish between the two is important, especially as the LLM revolution gathers speed.
It’s also worth remembering that the internet is a strange and often sinister place, and its darkest crevasses contain some of the raw material that’s training GPT-4 and similar AI tools. As Matteo detailed yesterday:
Microsoft’s original chatbot, named Tay and released in 2016, became misogynistic and racist, and was quickly discontinued. Last year, Meta’s BlenderBot AI rehashed anti-Semitic conspiracies, and soon after that, the company’s Galactica—a model intended to assist in writing scientific papers—was found to be prejudiced and prone to inventing information (Meta took it down within three days). GPT-2 displayed bias against women, queer people, and other demographic groups; GPT-3 said racist and sexist things; and ChatGPT was accused of making similarly toxic comments. OpenAI tried and failed to fix the problem each time. New Bing, which runs a version of GPT-4, has written its own share of disturbing and offensive text—teaching children ethnic slurs, promoting Nazi slogans, inventing scientific theories.
The latest in LLM tech is certainly clever, if debatably smart. What’s becoming clear is that those of us who opt to use these programs will need to be both.
Related:
ChatGPT changed everything. Now its follow-up is here. The difference between speaking and thinking Today’s News A federal judge in Texas heard a case that challenges the U.S. government’s approval of one of the drugs used for medication abortions. Credit Suisse’s stock price fell to a record low, prompting the Swiss National Bank to pledge financial support if necessary. General Mark Milley, the chair of the Joint Chiefs of Staff, said that the crash of a U.S. drone over the Black Sea resulted from a recent increase in “aggressive actions” by Russia. Dispatches The Weekly Planet: The Alaska oil project will be obsolete before it’s finished, Emma Marris writes. Up for Debate: Conor Friedersdorf argues that Stanford Law’s DEI dean handled a recent campus conflict incorrectly.Explore all of our newsletters here.
Evening Read Arsh Raziuddin / The AtlanticNora Ephron’s Revenge
By Sophie Gilbert
In the 40 years since Heartburn was published, there have been two distinct ways to read it. Nora Ephron’s 1983 novel is narrated by a food writer, Rachel Samstat, who discovers that her esteemed journalist husband is having an affair with Thelma Rice, “a fairly tall person with a neck as long as an arm and a nose as long as a thumb and you should see her legs, never mind her feet, which are sort of splayed.” Taken at face value, the book is a triumphant satire—of love; of Washington, D.C.; of therapy; of pompous columnists; of the kind of men who consider themselves exemplary partners but who leave their wives, seven months pregnant and with a toddler in tow, to navigate an airport while they idly buy magazines. (Putting aside infidelity for a moment, that was the part where I personally believed that Rachel’s marriage was past saving.)
Unfortunately, the people being satirized had some objections, which leads us to the second way to read Heartburn: as historical fact distorted through a vengeful lens, all the more salient for its smudges. Ephron, like Rachel, had indeed been married to a high-profile Washington journalist, the Watergate reporter Carl Bernstein. Bernstein, like Rachel’s husband—whom Ephron named Mark Feldman in what many guessed was an allusion to the real identity of Deep Throat—had indeed had an affair with a tall person (and a future Labour peer), Margaret Jay. Ephron, like Rachel, was heavily pregnant when she discovered the affair. And yet, in writing about what had happened to her, Ephron was cast as the villain by a media ecosystem outraged that someone dared to spill the secrets of its own, even as it dug up everyone else’s.
More From The Atlantic
“Financial regulation has a really deep problem” The strange intimacy of New York City Culture Break Colin Hutton / Apple TV+Read. Bootstrapped, by Alissa Quart, challenges our nation’s obsession with self-reliance.
Watch. The first episode of Ted Lasso’s third season, on AppleTV+.
P.S.“Everyone pretends. And everything is more than we can ever see of it.” Thus concludes the Atlantic contributor Ian Bogost’s 2012 meditation on the enduring legacy of the late British computer scientist Alan Turing. Ian’s story on Turing’s indomitable footprint is well worth revisiting this week.
— Kelli
Isabel Fattal contributed to this newsletter.
www.cnn.com › 2023 › 03 › 15 › tech › meta-ai-investment-priority › index.html
This story seems to be about:
www.cnn.com › 2023 › 03 › 14 › media › meta-layoffs-reliable-sources › index.html
This story seems to be about: