Friday, March 24, 2023

24: ChatGPT, GPT-4

AI Won’t Kill Our Jobs, It Will Kill Our Job Descriptions—and Leave Us Better Off
Meta’s New ChatGPT-Like AI Is Fluent in the Language of Proteins—and Has Already Modeled 700 Million of Them
Blueberries have joined green beans in this year’s Dirty Dozen list
The Operating Model of Tomorrow
Microsoft Researchers Claim GPT-4 Is Showing "Sparks" of AGI "We believe that GPT-4's intelligence signals a true paradigm shift in the field of computer science and beyond."
ELON MUSK APPARENTLY MADE A STUPID ENGINEERING DECISION THAT'S KNEECAPPED TESLA'S SELF-DRIVING PROMISES
Is GPT-4 Worth the Subscription? Here’s What You Should Know The new algorithm is exclusive to ChatGPT Plus, OpenAI’s subscriber-only service. We signed up and put it to the test.
A Radical Way of Thinking About MoneyMorgan Ricks argues that stricter regulations on “private money” can help panic-proof our financial system.
A Million Bodies Are Buried Here. Now It’s Becoming a Park. Hart Island, a potter’s field where New York City has buried its unclaimed dead for more than a century, will finally accept visitors this year.
Washington prepares for war with Amazon A multi-pronged investigation of the online giant is approaching the action phase — on mergers, antitrust, privacy and more.



ChatGPT as muse, not oracle
ChatGPT as muse, not oracle
ChatGPT Gets Its “Wolfram Superpowers”!
What Is ChatGPT Doing … and Why Does It Work?

It’s Just Adding One Word at a Time

....... applies just as well to other current “large language models” [LLMs] as to ChatGPT. ........ what ChatGPT is always fundamentally trying to do is to produce a “reasonable continuation” of whatever text it’s got so far, where by “reasonable” we mean “what one might expect someone to write after seeing what people have written on billions of webpages, etc.” .......... when ChatGPT does something like write an essay what it’s essentially doing is just asking over and over again “given the text so far, what should the next word be?”—and each time adding a word ........ it’s adding a “token”, which could be just a part of a word, which is why it can sometimes “make up new words” ....... there’s randomness here means that if we use the same prompt multiple times, we’re likely to get different essays each time ........ a simpler GPT-2 system, which has the nice feature that it’s small enough to be able to run on a standard desktop computer. ......... ChatGPT always picks its next word based on probabilities ........ There are about 40,000 reasonably commonly used words in English. ........ here’s the problem: there just isn’t even close to enough English text that’s ever been written to be able to deduce those probabilities. ......... In a crawl of the web there might be a few hundred billion words; in books that have been digitized there might be another hundred billion words. But with 40,000 common words, even the number of possible 2-grams is already 1.6 billion—and the number of possible 3-grams is 60 trillion. So there’s no way we can estimate the probabilities even for all of these from text that’s out there. And by the time we get to “essay fragments” of 20 words, the number of possibilities is larger than the number of particles in the universe, so in a sense they could never all be written down.
.

Will AIs Take All Our Jobs and End Human History—or Not? Well, It’s Complicated…
Wolfram|Alpha as the Way to Bring Computational Knowledge Superpowers to ChatGPT

Ukraine war: why China’s peace road map is the only one on the table State information campaigns are driving escalation, as are increased losses on both sides ...... Normalisation will require neutral arbiters who highlight common ground, such as that offered by trade in key commodities. Currently, China is the sole remaining neutral superpower ....... War is always a tragedy. Unfortunately, the war in Ukraine shows no signs of winding down. On the contrary, it is stuck in an escalatory spiral. Nato is supplying Ukraine with increasingly powerful weapons, including tanks, and Russia is not idly waiting for a Ukrainian offensive. ........... China offered an “off-ramp” in its peace proposal .......... early attempts at a settlement in March and April last year failed. ........

Beijing offered the only peace road map on the table

........ the risks of continuing down this path for China and Asia. ......... the information campaign that began in February 2022 makes normalisation a distant prospect. As most Western institutions – from governments to business and even academia – raced to break away from Russia, the prospect of finding common ground has been significantly reduced. ......... With streets and social media covered in Ukrainian flags, the political will needed to “explain away” the conflict becomes unattainable. Escalation remains the politically acceptable option and thus sets the direction. ........ Zelensky is pushing for increased military aid daily, using channels ranging from daily video statements to trying to address the Oscars. .......... Moscow’s open information campaign has relied mostly on RT and Sputnik, which enjoy growing reach in Africa and Latin America. Messaging via state media has also been intense domestically, positioning the conflict as an increasingly existential struggle. ............ As the level of sacrifice from both sides grows, peace becomes a more distant prospect. Just as in roulette,

it’s easier to go “double or nothing” than to quit with a loss

. ........... Nato leaders are sending heavier weapons in the hope that tanks, planes or missiles will eventually snap the front line. Russia has responded in kind with long-range weaponry. ........ there are no magic weapons – more guns just cost more lives. Ukrainian soldiers have spoken of mounting losses. Russian losses are growing too but simple military arithmetic remains in Russia’s favour. ......... The risk of further escalation is of earth-shattering proportions – ............ neutral arbiters that provide common ground. Russian and US foreign ministers met in India, not in Munich. ....... trade in key commodities like grain creates some common ground which can be a starting point for interaction. Beijing stressed the importance of keeping supply chains open, avoiding unilateral restrictions, and facilitating grain exports. ........ Trade difficulties in 2022 created inflation that had a global impact just as recovery from Covid-19 was under way. Further escalation will guarantee more macroeconomic shocks. ........... the world continues to slide down a slippery slope. Despite ongoing conventional support, the front line is stable, and sentiment in Russia mirrors Russian President Vladimir Putin’s statement last year that the army is “cracking [Nato weapons] like nuts”. ........ China is a neutral peacemaker. Russia is willing to seek compromise in line with Beijing’s proposals. The vector of development has been set, now the ball is in the West’s court. ........ Oleg Yanovsky is a lecturer in the Department of Political Theory at Moscow State Institute of International Relations (MGIMO).
.

AI Could Make More Work for Us, Instead of Simplifying Our Lives the introduction of automated processes that aim to simplify work—and free people’s time—can also make that work more complex, generating new tasks that many workers might perceive as mundane. ........ scientists were not released from repetitive, manual, or boring tasks as one might expect. Instead, the use of robotic platforms amplified and diversified the kinds of tasks researchers had to perform. ....... With automated methods, the possibilities are amplified. ....... had the effect of boosting the volume of data that needed checking, standardizing, and sharing. ........ Scientific work is often judged on output such as peer-reviewed publications and grants. However, the time taken to clean, troubleshoot, and supervise automated systems competes with the tasks traditionally rewarded in science. These less valued tasks may also be largely invisible—particularly because managers are the ones who would be unaware of mundane work due to not spending as much time in the lab. ........ challenges the assumption that everyone involved or affected by digitalization becomes more productive or has more free time when parts of their workflow are automated. ....... We also need to consider how to design and manage these processes so that technology can more positively add to human capabilities. .

OpenAI Says GPT-4 Is Better in Nearly Every Way. What Matters More Is Millions Will Use It In 2020, artificial intelligence company OpenAI stunned the tech world with its GPT-3 machine learning algorithm. After ingesting a broad slice of the internet, GPT-3 could generate writing that was hard to distinguish from text authored by a person, do basic math, write code, and even whip up simple web pages. ......... when it was founded as a nonprofit. (OpenAI reorganized as a capped-profit company in 2019.) ........ (Earlier this year, the free version of ChatGPT hit 100 million users faster than any app in history.) ........ It’s easy to imagine GPT-5 and other future models slotting into the ecosystem being built now as simply, and invisibly, as a smartphone operating system that upgrades overnight. ....... Chatbots in search, for example, may not outperform older approaches until they’ve proven to be far more reliable than they are today. And the cost of running generative AI, particularly at scale, is daunting. Can companies keep expenses under control, and will users find products compelling enough to vindicate the cost? ....... Indeed, companies are appending disclaimers warning users not to rely on them too much—like keeping your hands on the steering wheel of that Tesla. ....... millions of people may soon begin churning out presentation slides, emails, and websites in a jiffy, as the new crop of AI sidekicks arrives in force. .

ChatGPT Plug-Ins: OpenAI As A Platform Play

OpenAI turns ChatGPT into a platform overnight with addition of plugins and help ChatGPT access up-to-date information, run computations, or use third-party services.” .......

“I think the introduction of plugins to ChatGPT is a threat to the App Store. It creates a new platform with new monetization methods.”

........ OpenAI, he said, is offering a web browsing plugin and a code execution plugin..... The first plugins have already been created by companies including Expedia, Instacart, Kayak, OpenTable and Zapier. ......... According to Expedia, their new plugin simplifies trip planning for ChatGPT users. “Until now, ChatGPT could identify what to do and where to stay, but it couldn’t help travelers shop and book” ...... Now, once a traveler enables the Expedia plugin, they can bring a trip itinerary created through a conversation with ChatGPT “to life” with information powered by Expedia’s travel data including real-time availability and pricing of flights, hotels, vacation rentals, activities and car rentals. When ready to book, they’ll be sent to Expedia, where they can log in to see options personalized to what they prefer, as well as member discounts, loyalty rewards and more. .......... This is similar to how Apple’s App Store revolutionized the mobile industry by allowing third-party apps to flourish on its devices. ChatGPT’s plugin feature could potentially open up new possibilities and markets for AI chat in the future.




My Struggle Session at Stanford Law School A dean voices pride that students are being taught to stage tantrums rather than make a reasoned case........ Stanford Law School’s website touts its “collegial culture” in which “collaboration and the open exchange of ideas are essential to life and learning.” Then there’s the culture I experienced when I visited Stanford last week. I had been invited by the student chapter of the Federalist Society to discuss the U.S. Court of Appeals for the Fifth Circuit, on which I’ve served since 2018. I’ve spoken at law schools across the country, and I was glad to accept this invitation. One of my first clerks graduated from Stanford. I have friends on the faculty. I gave a talk there a few years ago and found it a warm and engaging place, but not this time. .......... When I arrived, the walls were festooned with posters denouncing me for crimes against women, gays, blacks and “trans people.” Plastered everywhere were photos of the students who had invited me and fliers declaring “You should be ASHAMED,” with the last word in large red capital letters and a horror-movie font. This didn’t seem “collegial.” Walking to the building where I would deliver my talk, I could hear loud chanting a good 50 yards away, reminiscent of a tent revival in its intensity. Some 100 students were massed outside the classroom as I entered, faces painted every color of the rainbow, waving signs and banners, jeering and stamping and howling. As I entered the classroom, one protester screamed: “We hope your daughters get raped!”

Diversity and Free Speech Can Coexist at Stanford We have to stop blaming, start listening, and ask ourselves: Is the juice worth the squeeze? ........ Stanford Law School’s chapter of the Federalist Society earlier this month invited Fifth Circuit Court of Appeals Judge Kyle Duncan to speak on campus. Student groups that vehemently opposed Judge Duncan’s prior advocacy and judicial decisions regarding same-sex marriage, immigration, trans people, abortion and other issues showed up to protest. Some protesters heckled the judge and peppered him with questions and comments. Judge Duncan answered in turn. Regardless of where you stand politically, none of this heated exchange was helpful for civil discourse or productive dialogue. ........ Students involved in the protest had previously requested that the event be canceled or moved to Zoom. In my role as Stanford Law School’s associate dean for diversity, equity and inclusion, I supported the administration’s decision not to cancel the event or move it to video, as it would censor or limit the free speech of Judge Duncan and the students who invited him. Instead, the administration and I welcomed Judge Duncan to speak while supporting the right of students to protest within the bounds of university policy.

Stanford Law Rediscovers Free Speech The dean instructs student hecklers on the First Amendment. ....... Stanford Law School disgraced itself two weeks ago when its diversity administrator let students heckle and shout down federal Judge Kyle Duncan. The school is now trying to salvage its reputation, and it’s making some progress....... In a letter to the university community on Wednesday, Stanford Law Dean Jenny Martinez issued a defense of free speech on campus and laid out the school’s expectations for civil discourse and legal professionalism.

Statement from Jenny Martinez As we consider the role of respectful treatment of members of our community, I want to be clear that the hate mail and appalling invective that have been directed at some of our students and law school administrators in the wake of March 9 are of great concern to me. All actionable threats that come to our attention will be investigated and addressed as the law permits. .

ChatGPT plugins We’ve implemented initial support for plugins in ChatGPT. Plugins are tools designed specifically for language models with safety as a core principle, and help ChatGPT access up-to-date information, run computations, or use third-party services. .......

Expedia, FiscalNote, Instacart, KAYAK, Klarna, Milo, OpenTable, Shopify, Slack, Speak, Wolfram, and Zapier.





The Age of AI has begun Artificial intelligence is as revolutionary as mobile phones and the Internet. ........ In my lifetime, I’ve seen two demonstrations of technology that struck me as revolutionary. ........ The first time was in 1980, when I was introduced to a graphical user interface—the forerunner of every modern operating system, including Windows. ....... The second big surprise came just last year. I’d been meeting with the team from OpenAI since 2016 and was impressed by their steady progress. In mid-2022, I was so excited about their work that I gave them a challenge: train an artificial intelligence to pass an Advanced Placement biology exam. Make it capable of answering questions that it hasn’t been specifically trained for. (I picked AP Bio because the test is more than a simple regurgitation of scientific facts—it asks you to think critically about biology.) If you can do that, I said, then you’ll have made a true breakthrough.........

I knew I had just seen the most important advance in technology since the graphical user interface.

........ The development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone. It will change the way people work, learn, travel, get health care, and communicate with each other. Entire industries will reorient around it. Businesses will distinguish themselves by how well they use it. ........ Globally, the worst inequity is in health ........ In the United States, the best opportunity for reducing inequity is to improve education, particularly making sure that students succeed at math. ....... Climate change is another issue where I’m convinced AI can make the world more equitable.

The injustice of climate change is that the people who are suffering the most—the world’s poorest—are also the ones who did the least to contribute to the problem.

............ it raises hard questions about the workforce, the legal system, privacy, bias, and more. AIs also make factual mistakes and experience hallucinations. ......... AGI doesn’t exist yet—there is a robust debate going on in the computing industry about how to create it, and whether it can even be created at all. ........ and they will get better very fast. ....... the early days of the personal computing revolution, when the software industry was so small that most of us could fit onstage at a conference. Today it is a global industry. ........... Soon the pre-AI period will seem as distant as the days when using a computer meant typing at a C:> prompt rather than tapping on a screen.......... AI will enhance your work—for example by helping with writing emails and managing your inbox. ....... Eventually your main way of controlling a computer will no longer be pointing and clicking or tapping on menus and dialogue boxes. Instead, you’ll be able to write a request in plain English. (And not just English—AIs will understand languages from around the world. In India earlier this year, I met with developers who are working on AIs that will understand many of the languages spoken there.) ......... a digital personal assistant: It will see your latest emails, know about the meetings you attend, read what you read, and read the things you don’t want to bother with. This will both improve your work on the tasks you want to do and free you from the ones you don’t want to do. ....... When productivity goes up, society benefits because people are freed up to do other things, at work and at home. ........ The rise of AI will free people up to do things that software never will—teaching, caring for patients, and supporting the elderly, for example. ........ The AI models used in poor countries will need to be trained on different diseases than in rich countries. They will need to work in different languages and factor in different challenges, such as patients who live very far from clinics or can’t afford to stop working if they get sick. ........ AIs will dramatically accelerate the rate of medical breakthroughs. The amount of data in biology is very large, and it’s hard for humans to keep track of all the ways that complex biological systems work.
.



"The hottest new programming language is English," Andrej Karpathy, Tesla's former chief of AI, said in a tweet in January.



140 reports of people trapped in lifts across parts of Hong Kong as suspected explosion at electricity pylon disrupts power supply Fire Services Department said it received at least 140 reports of residents ‘shut in lift’ Residents reported an explosion had occurred near Fei Ngo Shan, also known as Kowloon Peak, at around 11.30pm

Hong Kong must boost future investment and strengthen economy despite ‘fairly high’ fiscal deficit due to Covid-19 pandemic, finance chief says City’s fiscal reserve has shrunk following expansionary policies rolled out during pandemic, finance chief Paul Chan says ...... Treasury chief Christopher Hui stresses government ‘does not know how to grow money’ and must balance different factors to increase revenue

GPT-4 is OpenAI’s most advanced system, producing safer and more useful responses

Microsoft to Bring OpenAI’s Chatbot Technology to the Office Word, PowerPoint and Outlook emails will get new AI assistants called Copilots

Confirmed: the new Bing runs on OpenAI’s GPT-4
Interview with OpenAI’s Greg Brockman: GPT-4 isn’t perfect, but neither are you GPT-4 improves upon its predecessor, GPT-3, in key ways, for example giving more factually true statements and allowing developers to prescribe its style and behavior more easily. It’s also multimodal in the sense that it can understand images, allowing it to caption and even explain in detail the contents of a photo. ......... Like GPT-3, the model “hallucinates” facts and makes basic reasoning errors. In one example on OpenAI’s own blog, GPT-4 describes Elvis Presley as the “son of an actor.” (Neither of his parents were actors.) ....... On the AP Calculus BC exam, GPT-4 scores a 4 out of 5 while GPT-3 scores a 1. (GPT-3.5, the intermediate model between GPT-3 and GPT-4, also scores a 4.) And in a simulated bar exam, GPT-4 passes with a score around the top 10% of test takers; GPT-3.5’s score hovered around the bottom 10%. ........ GPT-4 can take a prompt of both images and text to perform some action (e.g. an image of giraffes in the Serengeti with the prompt “How many giraffes are shown here?”). ........ in internal tests, it was 82% less likely to respond to requests for content disallowed by OpenAI’s usage policy and 40% more likely to produce “factual” responses than GPT-3.5. ......... Bing Chat, Microsoft’s chatbot powered by GPT-4, has been shown to be highly susceptible to jailbreaking. Using carefully tailored inputs, users have been able to get the bot to profess love, threaten harm, defend the Holocaust and invent conspiracy theories. ........... OpenAI is testing a version of GPT-4 that can “remember” roughly 50 pages of content, or five times as much as the vanilla GPT-4 can hold in its “memory” and eight times as much as GPT-3........ “Previously, the model didn’t have any knowledge of who you are, what you’re interested in and so on,” Brockman said. “Having that kind of history [with the larger context window] is definitely going to make it more able … it’ll turbocharge what people can do.”

How Siri, Alexa and Google Assistant Lost the A.I. Race The virtual assistants had more than a decade to become indispensable. But they were hampered by clunky design and miscalculations, leaving room for chatbots to rise. ....... People have used ChatGPT to handle complex tasks like coding software, drafting business proposals and writing fiction. ........ large language models, which are systems trained to recognize and generate text based on enormous data sets scraped off the web ........... Google engineers spent years experimenting with its assistant to mimic what Alexa could do, including designing smart speakers and voice-controlled tablet screens to control home accessories like thermostats and light switches. The company later integrated ads into those home products, which did not become a major source of revenue. ........ Many of the big tech companies are now racing to come up with responses to ChatGPT.

Trump and DeSantis Could Both Lose
(Seven of the 10 fastest growing states have Republican governors while eight of the 10 fastest shrinking states have Democratic governors.) ........ Gov. Brian Kemp, for example, is making Georgia a hub for green manufacturing, attracting immense investments in electric vehicle technologies. In his inaugural address he vowed to make Georgia “the electric mobility capital of America.” As Alexander Burns noted in Politico, Kemp doesn’t sell this as climate change activism; it’s jobs and prosperity. ....... The Donald Trump/Tucker Carlson orbit is rife with indignation and fury. ........ The Florida governor should be the ultimate optimistic, businesslike conservative. His state is growing faster than any other in the country. But instead, he’s running as a dour, humorless culture war populist — presumably because that’s what he is. .

China’s Answer to ChatGPT Gets an Artificial Debut and Disappoints The promised “live” demonstration of the bot had, in fact, been recorded. Shares of Baidu, the company behind the technology, tumbled in Hong Kong......... Halfway through a demonstration that had been marketed as live, in which Ernie summarized a science fiction novel and analyzed a Chinese idiom, Robin Li, Baidu’s chief executive, said the presentation had been prerecorded “to save time.” ......... Baidu’s shares plunged 10 percent in Hong Kong, a striking contrast to the rally this year that was fueled by the company’s announcement that it has had a rival to ChatGPT in the works since 2019. ........ As Washington has moved to contain competition from China, it has cut Beijing off from high-end computing chips — a key ingredient in technologies like ChatGPT and Ernie. ....... “China is incredibly good at scaling an existing invention, but it is not very good at making breakthroughs” ........

The country ... lacks the diversity of thought and free expression of ideas that help nurture out-of-the-box thinking.

.

The Silicon Valley Bank Rescue Just Changed Capitalism Now, in light of the bank failures of the last few days and the F.D.I.C.’s extension of coverage, why will any depositor worry about risk? Having bailed out depositors of two banks in full, how will the government refuse others? ....... In 1933, an estimated 4,000 banks failed ........ Banks like Lehman had too much leverage, and they were overexposed to a very weak and widely held asset, mortgage securities. ......... The mismatch between the cost of their money and the (lower) rate that their mortgages earned sank the industry. ....... The first plank of capitalism is that it entails risk ...... In the case of Signature, which was exposed to the crypto industry, the rescue probably bailed out gamblers on speculative assets. ........ Most likely, banks will pass along the rescue costs in the form of higher fees to consumers. ........ Strictly speaking, President Biden’s assurance that taxpayers are not on the line was accurate. However, in the sense that banking customers are a pretty big group, the “public” will be affected. ........ The regulators clearly failed to monitor S.V.B.’s unhealthy mismatch of assets and liabilities. Their job will be more difficult in the future, as risk taking on deposits has effectively become socialized. What if a bank opts to attract more funds by raising its interest rate on deposits? Can the regulators permit it? Wait a second, this is what all banks do. .

10 Ways GPT-4 Is Impressive but Still Flawed OpenAI has upgraded the technology that powers its online chatbot in notable ways. It’s more accurate, but it still makes things up.......... It is an expert on an array of subjects, even wowing doctors with its medical advice. It can describe images, and it’s close to telling jokes that are almost funny. ....... the system could describe an image from the Hubble Space Telescope in painstaking detail. The description went on for paragraphs. ........ it can exhibit this kind of expertise across many areas, from computer programming to accounting. ........ It can also score a 1,300 (out of 1,600) on the SAT and a five (out of five) on Advanced Placement high school exams in biology, calculus, macroeconomics, psychology, statistics and history .......... Though the new bot seemed to reason about things that have already happened, it was less adept when asked to form hypotheses about the future. It seemed to draw on what others have said instead of creating new guesses. ......... The new bot still makes stuff up. Called “hallucination,” the problem haunts all the leading chatbots. Because the systems do not have an understanding of what is true and what is not, they may generate text that is completely false. .