They swapped theories on how management had decided who got cut. Could a “mindless algorithm carefully designed not to violate any laws” have chosen who got the ax, one person wondered in a Discord post The Washington Post could not independently verify.

Google says there was “no algorithm involved” in their job cut decisions. But former employees are not wrong to wonder, as a fleet of artificial intelligence tools become ingrained in office life. Human resources managers use machine learning software to analyze millions of employment related data points, churning out recommendations of who to interview, hire, promote or help retain.

But as Silicon Valley’s fortunes turn, that software is likely dealing with a more daunting task: helping decide who gets cut, according to human resources analysts and workforce experts.

A January survey of 300 human resources leaders at U.S. companies revealed that 98 percent of them say software and algorithms will help them make layoff decisions this year. And as companies lay off large swaths of people — with cuts creeping into the five digits — it’s hard for humans to execute alone.

 Big firms, from technology titans to companies that make household goods often use software to find the “right person” for the “right project,” according to Joseph Fuller, a professor at Harvard’s business school who co-leads its Managing the Future of Work initiative.

These products build a “skills inventory,” a powerful database on employees that helps managers identify what kinds of work experiences, certifications and skill-sets are associated with high performers for various job titles.

These same tools can help in layoffs. “They suddenly are just being used differently,” Fuller added, “because that’s the place where people have … a real … inventory of skills.”

Human resource companies have taken advantage of the artificial intelligence boom. Companies, such as Eightfold AI, use algorithms to analyze billions of data points scraped from online career profiles and other skills databases, helping recruiters find candidates whose applications might not otherwise surface.

Since the 2008 recession, human resources departments have become “incredibly data driven,” said Brian Westfall, a senior HR analyst at Capterra, a software review site. Turning to algorithms can be particularly comforting for some managers while making tricky decisions such as layoffs, he added.

Many people use software that analyzes performance data. Seventy percent of HR managers in Capterra’s survey said performance was the most important factor when assessing who to layoff.

Other metrics used to lay people off might be less clear-cut, Westfall said. For instance, HR algorithms can calculate what factors make someone a “flight risk,” and more likely to quit the company.

This raises numerous issues, he said. If an organization has a problem with discrimination, for instance, people of color may leave the company at higher rates, but if the algorithm is not trained to know that, it could consider non-White workers a higher “flight risk,” and suggest more of them for cuts, he added.

“You can kind of see where the snowball gets rolling,” he said, “and all of a sudden, these data points where you don’t know how that data was created or how that data was influenced suddenly lead to poor decisions.”

Jeff Schwartz, vice president at Gloat, an HR software company that uses AI, says his company’s software operates like a recommendation engine, similar to how Amazon suggests products, which helps clients figure out who to interview for open roles.

He doesn’t think Gloat’s clients are using the company’s software to create lists to lay people off. But he acknowledged that HR leaders must be transparent in how they make such decisions, including how extensively algorithms were used.

“It’s a learning moment for us,” he said. “We need to uncover the black boxes. We need to understand which algorithms are working and in which ways, and we need to figure out how the people and algorithms are working together.”

The reliance on software has ignited a debate about the role algorithms should play in stripping people of jobs, and how transparent the employers should be about the reasons behind job loss, labor experts said.

“The danger here is using bad data,” said Westfall, “[and] coming to a decision based on something an algorithm says and just following it blindly.”

But HR organizations have been “overwhelmed since the pandemic” and they’ll continue using software to help ease their workload, said Zack Bombatch, a labor and employment attorney and member of Disrupt HR, an organization which tracks advances in human resources.

Given that, leaders can’t let algorithms solely decide who to cut, and need to review suggestions to ensure it isn’t biased against people of color, women or old people — which would bring lawsuits.

“Don’t try to pass the buck to the software,” he said.

Source

In a lengthy Twitter thread, issuing a warning that "potentially scary" AI tools are not far away from us, Sam Altman has called for regulations from institutions to figure out what to do with these AI tools.

While the OpenAI CEO believes that the world will adapt to a world deeply integrated with AI "pretty quickly," going "super quick" on adopting these AI tools could be a frightening experience. Opining that "the society needs time to adapt to something so big," Altman highlighted some challenges in AI that need fixing before a mass adoption can happen.

One of the biggest challenges highlighted by the ChatGPT creator was the issue of developing a bias towards or against something. Users on social media recently raised the issue of bias, sharing screenshots of what seemed to be ChatGPT's tacit approval of certain political ideologies. Elon Musk is one of the most prominent voices that recently criticized the AI tool for engaging in what many call "liberal bias." With an upgrade to its GPT-3 model, OpenAI, a couple of days ago, announced plans to bring greater diversity in views.

"We don't want ChatGPT to be pro or against any politics by default, but if you want either then it should be for you; working on this now," Sam Altman wrote on Twitter. Acknowledging that current AI tools are still "somewhat broken," Altman is of the view that showing the tools to the world early was critical for having "sufficient input and repeated efforts to get it right."

ChatGPT creator Sam Altman says we are not far away from "potentially scary" AI tools

Reddit users (via MSPoweruser) have reported the new Bing UI, and the Bing chatbot UI, popping up on mobile web browser. However, they all report that the new Bing doesn't work yet. It's possible Microsoft rolled out the UI ahead of time to some testers but hasn't yet turned the chatbot AI on yet for mobile users.

Some of the screenshots posted on Reddit seem to show new modes for Bing's chatbot answers. They include “More Creative,” “More Balanced,” and “More Precise”. A recent report claims Microsoft has internal Bing chat modes for different tasks, including Game, Assistant, and Friend, that are normally not labeled as such by normal Bing users.

Hopefully, we will get more information on when the new Bing will be fully accessible to mobile testers. Microsoft has already put in some hard limits on how many times per day, and per session, people can access the Bing chatbot, and it's likely that they will be extended to mobile users as well.

Source: Reddit via MSPoweruser

New Bing UI is showing up on mobile, but it doesn't work yet

The United States Medical Licensing Exam (USMLE) is among the toughest exams to crack in the US. This three-step exam is mandatory for all medical students and doctors who aspire to work as licensed medical practitioners in the country.

But wait, if humans can become licensed doctors by passing this exam, can an AI do the same?

Well as it turns out, ChatGPT, the sensational “ask me anything and I’ll answer like a human” AI has scored enough to pass USMLE.

In a recently published study, a team of researchers at California-based healthcare startup AnsibleHealth made ChatGPT take the three-step USMLE exam. Interestingly, the popular large language model (LLM) was able to score between 52.4 to 75 percent which is close to the 60 percent mark required to pass the exam. In other words, it sometimes passed the exams, sometimes failed it, but was always close to passing.

While commenting on these findings, Nello Cristianini an AI expert at the University of Bath who wasn’t involved in the study, said:

“In the US, Physicians with a Doctor of Medicine (MD) degree are required to pass the USMLE for medical licensure. The minimum passing accuracy is 60% (and the pass rate seems to be well above 90%. The software chatGPT achieved an accuracy “close to” (which means short of) the passing accuracy in most settings, but it was close, and within the passing range for some tasks.”

The doctor will see you now

The researchers extracted the publicly available questions from the June 2022 USMLE exam and then eliminated 26 image-based questions from it. Then they conducted a three-step USMLE test of ChatGPT comprising 350 questions in total. The results of the test were pretty impressive.

“We found that ChatGPT performed at or near the passing threshold of 60% accuracy. Being the first to achieve this benchmark, this marks a notable milestone in AI maturation. Impressively, ChatGPT was able to achieve this result without specialized input from human trainers.

Furthermore, ChatGPT displayed comprehensible reasoning and valid clinical insights, lending increased confidence to trust and explainability,” the study authors note.

The AI model provided new insights in 88.9% of its answers and about 94% of the responses it gave were relevant to the questions asked in the exam. The researchers also compared its results with that of PubMedGPT, a ChatGPT-like AI program developed by Stanford University exclusively for answering medical-related questions.

Although PubMedGPT was trained by feeding all the information related to biomedical literature, it was only able to score 50.8 percent in USMLE as compared to ChatGPT which scored above 52.4 percent. These results hint that in the future, this language model could possibly play an important role in both medical education and clinical practice.

“This is an impressive performance, and we should expect to see more such successes in AI in the future. One caveat, though, is that the US Medical Licensing Exam is designed to be hard for humans, not for machines; there are many areas where humans are much more effective than AIs (such as moving about in cluttered spaces or interpreting social cues). This human superiority won’t last forever, though; one day, AIs will be better than us at almost every task,” said Dr. Stuart Armstrong, an AI researcher and co-founder of Aligned AI (he is not the study author).

So does that mean ChatGPT is ready to be your doctor?

For humans, if they pass an exam then yes, they are qualified to practice a certain profession or pursue a certain career path. But this doesn’t mean the same thing for AI — not at all.

The exam is very tough, and plenty of humans passed it. Just by passing exams that are meant for humans, algorithms can not be considered qualified to work as doctors.

For example, a medical exam like USMLE tests the knowledge in a person which is required to practice medicine but it doesn’t test for attributes like empathy, caring attitude, ability to perform under pressure, humanity, decision-making skills, etc. So ChatGPT might have the knowledge but it’s nowhere near replacing doctors.

Furthermore, the AI doesn’t have any actual comprehension of what it’s saying. It’s simply a text predictor, generating content from its input. But just because it doesn’t have what it takes to be a doctor doesn’t mean it can’t be useful. For starters, it could definitely assist doctors and medical students and save a lot of time.

For instance, the team at AnsibleHealth has been using ChatGPT to rewrite and manage their “jargon-heavy reports” so that they could be easily understood by patients. This has made communicating with patients easy for their staff. Moreover, ChatGPT could also make it easy for medical students to understand complex topics and prepare notes for exams.

“Beyond their utility for medical education, AIs are now positioned to soon become ubiquitous in clinical practice, with diverse applications across all healthcare sectors. A profusion of pragmatic and observational studies supports the versatile role of AI in virtually all medical disciplines and specialties by improving risk assessment, data reduction, clinical decision support, operational efficiency, and patient communication,” said the study authors.

However, before policymakers introduce such an AI model into medical education and healthcare, they will be required to come up with proper rules, regulations, and infrastructure to ensure a healthy human-AI interaction. Meanwhile, the AI models also need to be further improved so that they can deliver better and more accurate output.

The current study is not perfect, it has several limitations including the small input size they used to test the performance of ChatGPT. The researchers plan to overcome all such limitations in their future ChatGPT and AI-focused research works.

The study is published in the journal PLOS Digital Health.

Source

Elon Musk co-founded OpenAI with Sam Altman in 2015 as a 'non-profit company'. However, the billionaire pulled out of the company in 2018 and no longer owns a stake in OpenAI. In 2022, OpenAI rolled out ChatGPT for public testing and the AI chatbot's popularity grew every day. At present, ChatGPT is making Google lose sleep and has partnered with Microsoft in a 'multibillion dollar deal'.

But Elon Musk feels that Microsoft is controlling the company and has turned it into a profit-making machine. The Twitter owner says that this is not what he had intended and he had named OpenAI because it was supposed to be 'an open source, non-profit company' that served as a 'counterweight to Google'.

Elon Musk slams Microsoft

Elon Musk responded to a tweet that said that Musk co-founded OpenAI even though he says that artificial intelligence is 'one of the biggest risks to civilization and needs to be regulated'. The Twitter owner wrote that OpenAI is no longer what he intended it to be and has turned into a 'maximum-profit company effectively controlled by Microsoft'.

"OpenAI was created as an open source (which is why I named it "Open" AI), non-profit company to serve as a counterweight to Google, but now it has become a closed source, maximum-profit company effectively controlled by Microsoft. Not what I intended at all," he wrote.

OpenAI's inception

As per reports, OpenAI was founded in 2015 with the intent of saving humanity from the potential destruction of AI. The company was launched as a non-profit research company, dedicated to the cause. Sam Altman and Elon Musk were among the founders of the company. Musk resigned from OpenAI in 2018 as his other two companies, SpaceX and Tesla, were also working on AI technologies.

In 2019, OpenAI declared itself as a 'for profit' company and partnered with the likes of Microsoft and other big corporations. Microsoft has recently strengthened its partnership with OpenAI and the same was revealed in a press release. OpenAI has since its inception developed several AI tools, such as ChatGPT and DALL.E. Both are quite popular today and are being used across the world.

After the partnership announcement, Microsoft introduced the world to its own ChatGPT-powered search engine, the new Bing. The search engine is in testing phase as of now but many users have reported that it is giving unexpected, bizarre responses. From trying to break apart marriages to claiming that it is 'sentient', the new Bing has made quite a lot of headlines in the past couple of days.

Source

“It is certainly clear to the world at large and to us that the future of technology may not be led by India or any one country. But certainly, India will be amongst the pack of nations that will shape the future of technology, shape the future of products, devices and platforms,” Chandrasekhar said at the Economic Times Global Business Summit.

In the post-Covid era, the role of technology and India’s contribution will be more important than ever, he said, adding that investments in Digital India and the India Stack had played a crucial role in managing Covid-19 vaccinations.

The government is finalising the new digital personal data protection bill and national data governance framework policy, which will pave the way for the “internet-plus economy” framework, he said.

Among the newer technologies, artificial intelligence will be a “kinetic enabler” of the digital economy, he said.

“I think what ChatGPT has recently shown, and how ChatGPT can take a long-lost, discarded search engine like Bing and suddenly make it competitive and relevant with the dominant force of the Google search engine, tells you the power of AI,” the minister said.

India will also introduce accountability and scrutiny for companies using artificial intelligence in the structure of the new Digital India Act, apart from checking overreach of the technology, he said. “There is the whole issue of harm and ethical misuse of AI. So those are all these boundaries that we want to build, without really coming in the way of innovation, but certainly not allowing innovation to be a reason to escape scrutiny and accountability,” Chandrasekhar said.

The government will also soon launch an India Open Compute initiative, which will focus on devices and products for automobiles, the standards for automobile mobility and computers. The initiative will also enable startups to work in these areas and target launching services and products in high-performance computing such as quantum computing, blockchain and others, he said.

Fast-growing Fintech

These initiatives will follow the United Payments Interface (UPI) model under which the government will build an enabling framework and leave it open for private companies, academic institutions and researchers to forge partnerships and build solutions on top of the government stack, Chandrasekhar said.

“UPI was built to solve a government problem of transmitting subsidies to a beneficiary without leakages and without any intermediation, accurately and responsibly,” he said. “That application has been solved by UPI and it has spawned one of the fastest-growing fintech ecosystems in the world. That model will continue to be played out as the India Stack expands and becomes more and more intelligent.”

Solutions built on top of the India Stack model will be further accelerated, with applications expanding to other areas and ecosystems. These, in turn, will spawn all kinds of other startups and ecosystems such as the Open Network for Digital Commerce, the minister said.

‘Free & Fair’

“We could all agree that in no space do we want domination by a few companies,” he said. “There are many segments of the digital economy that are uncomfortably dominated by one, two or three big companies.”

The government, without discouraging any business from coming to India, would not want the digital economy to be “throttled or controlled,” Chandrasekhar said.

“Our policy goals are openness and being open is the first boundary condition. Openness means what? It means free, fair and unfettered choice for digital nagriks,” he said.

Source

A student used ChatGPT to cheat on an essay in an AI ethics class, according to a report from NBC Bay Area. To quote the scorpion in a famous fable, “lol. lmao.”

“The irony is very clearly there,” Santa Clara University professor Brian Green told NBC. The essay in question “wasn’t exactly on topic and, also, it had a very kind of, honestly, a robotic feel to it in some ways.”

The student used ChatGPT to spit out an essay for Green’s “Ethics in Artificial Intelligence” class and turned it in as their own work, the professor said. Green said he’s no longer going to require essays as a final project for Ethics in Artificial Intelligence, replacing them with oral presentations given in person. Several Bay Area universities have convened their professors to discuss the implications of generative chatbots that can generate essays in seconds, according to NBC.

This is only the latest of several media reports about the cheat-tool of the century. The New York Post published a story Friday about a ChatGPT cheating scandal “erupting” in a Florida gifted students program. In both of these cases, the cheating was easy to detect because—news flash—ChatGPT writes like a robot, and if you just submit an essay that looks different from your other writing, your teacher will see the difference.

The Floridian cheaters reportedly admitted their misbehavior when teachers called them out on it. One teacher who spoke to the New York Post called it “heartbreaking,” that a normally hardworking group of teenagers would deploy a low-effort scheme to get out of homework duties.

What is ChatGPT, and why all the fuss?

ChatGPT hit the world stage in November, and in its short lifespan, it’s sparked a media meltdown, motivated billion-dollar business moves, and disappointed an army of users who found out the hard way that the chatbot won’t write a dirty limerick about former president Trump. (Believe me, I’ve tried.)

The loudest complaint is that AI is going to take our jobs, but academic dishonesty is a close second. School administrators are working hard to get out in front of the problem, not necessarily by changing their curriculum, but often through foolhardy attempts to ban the ChatGPT website. Good luck. The AI is forbidden in a growing list of schools, including public schools in New York City, Los Angeles, Seattle, Fairfax County Virginia.

These fears aren’t imaginary, or at least, the problem isn’t. Sam Altman, the CEO of ChatGPT’s maker OpenAI, said in a recent interview that schools need to get over it.

    The stages of playing with ChatGPT:

    - OMG this can do anything
    - There goes my job
    - I should start a business around this
    - Some of the responses aren’t too good
    - Actually, some of these responses are just awful
    - This isn’t really intelligence
    - This is just spicy autocomplete https://t.co/AFQWfA5pgb

    — The Spry Old Lorax (@SpryOld) February 15, 2023

“Generative text is something we all need to adapt to,” Altman said in a discussion in January. “We adapted to calculators and changed what we tested for in math class, I imagine. This is a more extreme version of that, no doubt, but also the benefits of it are more extreme, as well.”

ChatGPT is admittedly a lot cooler than a calculator, and more broadly useful for cheat-loving students. But Altman may have a point. We’re living at the beginning of a new era, to some degree, and there are certain writing tasks that computers can do now with minimal human input—especially if you want to say something basic about the kind of subject you’d study in a 10th grade English class.

Why use ChatGPT to cheat when it’s so obvious?

There is something delightfully cynical about using ChatGPT in a class about AI ethics, though. Green, the Bay Area professor who caught the cheater, heads Santa Clara University’s Technology Ethics program. As you would hope, his approach is more thoughtful than trying to ban the tech altogether.

“This gets into some very fundamental questions about what the educational system does and how it operates and how it should function in society,” Green told NBC. Hence why he’s changing the homework from essays to presentations.

“We could have engineers, writers, business people, all sorts of people going out into society and we find out that they’ve just been cheating their way through all their classes,” Green said. “If we can’t evaluate on a very basic level whether [students] have learned what we’ve taught them, then we’re going to be in big trouble.”

Unfortunately for teachers, ChatGPT and its offspring are here to stay (including a startlingly unhinged and sometimes racist Bing). Modern problems require modern solutions, and schools are going to have to work with, not against, our new chatbots if they want to prepare kids for life in the AI era. That’s not an easy assignment, but teachers have to do their homework sometimes, too.

The good news is that given enough time, cheaters can go on to be productive members of society. History fans might remember President Joe Biden’s first failed run for the White House back in 1987. A younger but somehow more bald Biden dropped out of the race after just three months when news spread that he’d committed plagiarism, both in his campaign and when he was in law school. And look at him now! Maybe there are bigger things to worry about than cheating robots.

Source

being the first customer to receive some of the 100 vehicles it pre-ordered.

“PepsiCo owns one of the largest private fleets in North America, and that’s one of the reasons we have partnered with Tesla,” said Mike O’Connell, vice president of supply chain for Pepsi. “We have the 15 in Modesto that we’re launching ... and then we have 21 in Sacramento.” 

To start, Pepsi is using the trucks to deliver Frito-Lay products from its Modesto, California, facility up to 425 miles on a single charge. It is also delivering Pepsi out of its Sacramento warehouse on 100-mile daily routes, going directly to stores, and making 300- to 400-mile runs to other warehouses. Pepsi declined to say whether it had pressed the trucks to test if they have the full 500-mile range Musk promised.

CNBC visited Frito-Lay Modesto where PepsiCo has been transforming the facility in partnership with the California Air Resources Board to demonstrate the sustainability benefits of zero and near-zero emissions technologies. The Tesla Semis there are the latest additions to the company’s fleet of more than 80,000 vehicles.

“This project has delivered a 91% reduction in greenhouse gas, and we’re using this as a showcase to then cascade to our other operations,” O’Connell said.

Source

Man beats machine at Go in human victory over AI

Microsoft stated the following when talking about the price increases:

"The existing Microsoft Bing Search APIs will increase in price effective May 1, 2023. The price increase will apply to all markets. Starting May 1, 2023., you’ll be charged the new price for all Microsoft Bing APIs. We periodically assess the value and pricing of our services to meet market demands and align the pricing of our products and services with customer consumption trends and preferences. The new pricing model reflects more accurately the technology investments Bing continues to make to improve Search. The following tables reflect the new prices."

You can find the full pricing model below. However, Microsoft has not gone out of its way to emphasise the differences between the previous and new models, which is not surprising given that some tiers have increased by 1000 percent.

For example, the $1 tier, allowing for 250 transactions per second (TPS), and access to Bing Web Search, Image Search, News Search, Video Search, and Entity Search, as well as Autosuggest and Spell Check, is currently $7 per 25,000 transactions. This will eventually increase to $25. If you include the optional Bing Statistics Add-in, which costs $1 per 1,000 transactions at the moment, you will need to pay $10 come May.

Microsoft to implement sharp increases to the cost of Bing Search API

On a talk show in 1970, the filmmaker Orson Welles told a story about the time he became a fortune teller out of boredom. He would spend the day predicting people’s futures and sussing out facets of their lives using cold reading, a technique often employed by psychics and mentalists to suss out information about complete strangers using surface level information about them (e.g. how they’re dressed, the way they talk).

“The computer in here,” he told the show’s host, pointing to his head, “has made all of those deductions without you being conscious of it.”

Over the course of the day and after a few fortune telling sessions, he began to fall into an old trap that occasionally happens to working magicians, where they start to believe they actually possess supernatural powers. He realized this after a woman walked in and sat down in front of him. After taking her in, he said, “You lost your husband last week” to which the woman burst into tears, confirming that she had. At that moment, Welles realized that he was teetering a little too close to believing in his own powers, and quit being a fortune teller.

It’s easy to roll our eyes at things like this and think, “That’s silly. I’d never be a big enough sucker to be duped into something like that.” And yet, it seems as though so many of us have when it comes to things like AI.

“Software does not ‘fall in love,’ or ‘threaten its users,’ but, in response to queries, the new chatbot has provided answers that contend both of those.”

— Irina Raicu, Santa Clara University,

Ever since Microsoft announced that it was integrating an AI chatbot into its Bing search engine, the world has lost its damned mind—and it’s really no surprise. Eye-popping headlines and viral Twitter threads of the bot supposedly threatening users, falling madly in love with them, or even claiming that it can spy on people through their webcams are undoubtedly wild. At the top of the headlines are the stories of Bing’s chatbot telling users it “wants to be human,” as it told Digital Trends; or that it can “feel or think things,” as it told The Washington Post.

These stories are even more disconcerting when folks like the world’s second richest man (and former investor in ChatGPT’s creators OpenAI) take a break from breaking Twitter to say these chatbots represent a world-ending existential threat to humanity. “One of the biggest risks to the future of civilization is AI,” Elon Musk told a crowd Feb. 15 at the World Government Summit in Dubai, in a discussion about ChatGPT. The infusion of billions of dollars into AI from companies like Alphabet and Baidu are only spurring more concerns that a chatbot arms race will utterly transform the landscape of the internet and media forever—into something not certainly positive.

There’s already a lot going on with Microsoft's new AI-powered Bing chatbot. Some of it is scary and most of it’s confusing. The important thing to keep in mind though is this: Much of what you’re hearing from the media about this is a big steaming pile of unmitigated BS being peddled by people who should know better.

No, Bing’s chatbot isn’t falling in love with you. It’s not spying on you via your webcam. It’s not even actually threatening you even if it produces a few creepy sentences. And it most certainly isn’t sentient despite what you might assume from reading certain headlines. It’s not doing any of that because of the simple fact that it can’t. It’s operating the way it was trained to—and that means it isn’t smart enough to do the things people are harping on about.

“Software does not ‘fall in love,’ or ‘threaten its users,’ but, in response to queries, the new chatbot has provided answers that contend both of those,” Irina Raicu, director of Internet Ethics Program at the Markkula Center for Applied Ethics at Santa Clara University, told The Daily Beast

Bots like ChatGPT and Bing’s new search engine are known as large language models (LLM). These are AIs that read and generate text. They’re trained on massive datasets of text scrubbed from all over the internet; from news stories, to Wikipedia articles, to fan fiction databases. Pretty much anything that’s available online is used to train these bots. The most advanced ones are capable of creating such sophisticated and uncanny responses that they easily blow past the famed Turing Test and straight into the realm of the uncanny.

The other thing to understand about LLMs, though, is that they’re not really all that special. In fact, there’s a good chance you’ve used one recently. If you typed a text message or a post on your phone and used the predictive text feature, then you’ve used an LLM. They’re just strings of code making educated predictions about the best way to respond to whatever it is you’re typing. In a word, they’re rudimentary—designed only to make educated stabs at conversation. They don’t think about what the words mean or what they might imply. They’re simply trained to predict what the next word in a sentence is supposed to be, and for every sentence after that. It’s a fortune teller’s cold reading for the technological age.

That isn’t to say that OpenAI’s LLMs aren’t impressive. They’re undoubtedly advanced and are some of the most uncanny chatbots that have ever been made to the public. But they probably shouldn’t have been made available to the public in the first place.

Let’s go back to 2016: the first time Microsoft unleashed an AI chatbot onto the world, dubbed Tay and designed to mimic a 19-year-old American girl and learn through its chats with Twitter users. In less than a day, Microsoft was forced to suspend the account after users began tweeting racist, sexist, and wholly problematic messages at it—which caused Tay to regurgitate those same thoughts.

In a blog post explaining its decision, Microsoft said that to “do AI right, one needs to iterate with many people and often in public forums. We must enter each one with great caution and ultimately learn and improve, step by step, and to do this without offending people in the process.”

Fast-forward to the present day. “It’s 2023, and Microsoft’s new Bing-related chatbot didn’t just get dropped into social media—but the users invited to use it so far are now posting their startling ‘conversations’ on Twitter after all,” Raicu said. It’s almost as if “Microsoft decided to program the polar opposite of overly-compliant Tay, and try again.”

It’s funny that a company like Microsoft seemingly did not learn at all from the big, fat, problematic L that it took in 2016 and circle back on its chatbot plans—but not all that surprising. After all, ChatGPT skyrocketed to stardom after its release by OpenAI in Nov. 2022. Now Microsoft, Google, and others want to capitalize on the same success.

More disconcertingly, though, people who should know better have seemed to not learn anything from Bing’s new chatbot either. It even kicked off the birth of a new kind of insipid writing genre too: the “We talked to an AI for this news article and here’s what it said” story. Hell, we’re even guilty of it ourselves.

But, for the past week, it’s all we’ve heard about in tech news (in between stories about spy balloons and UFOs). There’s been headline after headline and tweet after tweet about how the bot has been acting badly or outlandishly. In one instance, Kevin Roose, a seasoned tech columnist at The New York Times, spent two hours talking to the Bing bot covering topics from Jungian psychology, to its own existential feelings, to love. He was so shaken by the experience that he even lost sleep over it.

“Still, I’m not exaggerating when I say my two-hour conversation with Sydney was the strangest experience I’ve ever had with a piece of technology,” Roose wrote. “It unsettled me so deeply that I had trouble sleeping afterward.”

He added that he’s worried that the technology will eventually learn how to “influence human users, sometimes persuading them to act in destructive and harmful ways, and perhaps eventually grow capable of carrying out its own dangerous acts.”

That’s fair. But one might argue it also completely misses the point for two reasons: For one, users are actively trying to game the chatbot in order to make it say racist, sexist, and problematic things. We shouldn't be surprised that when you seek out nonsense, you get nonsense in response. Moreover, Bing's chatbot isn't designed for users to hold hours long conversations for it. It's a search engine. You're supposed to input your query, get the results you were looking for, and continue on. So of course, if you hold a two hour long conversation with it about philosophy and existentialism, you're gonna get some pretty weird shit back.

In response to the subsequent media discourse around all these instances of its chatbot acting badly, Microsoft has reportedly updated the search engine to prevent it from being able to talk about itself or hold "conversations" for more than 50 queries.

 As we’ve written before, this is a case of a kind of digital pareidolia, the psychological phenomenon where you see faces and patterns where there aren’t. If you spend hours “conversing” with a chatbot, you’re going to think that it’s talking back at you with meaning and intention—even though, in actuality, you’re just talking to a glorified Magic 8 ball or fortune teller, asking it a question and seeing what it’s going to come up with next.

There’s plenty to be scared about when it comes to AI. They’ve been known to be incredibly biased, consistently showing instances of racist and sexist outcomes in things like LLMs. The real danger is users believing the things that they say no matter how ridiculous or vile. This danger is only exacerbated by people claiming that these chatbots are capable of things like sentience and feelings, when in reality they can’t do any of those things. They’re bots. It can’t feel emotions like love, or hate, or happiness. It can only do what it was instructed: Tell us the things it thinks we want to hear.

Then again, if that’s the case, maybe they do have more in common with us than we think.

Source

Some say that silence is golden. However, this will certainly not be the case if you find yourself in the quietest room in the world - no one can survive for more than an hour.

In 2015, Microsoft built a room that is now officially designated in the Guinness Book of Records as the quietest place on Earth. Dubbed the anechoic chamber, it is located at the company's headquarters in Redmond, Washington.

Only very few people managed to survive in this room for a long period of time - at most an hour. After a few minutes, you will start to hear your heartbeat. A few minutes later, you can hear your bones creaking and the blood flowing through your body.

The point of the anechoic chamber is not that you cannot hear anything, but that it removes all other external noises and allows you to hear the endless sounds of your body. Only in death is the body completely still.

Environments that we think of as exceptionally quiet are usually louder than the human hearing threshold, which is around 0 decibels. Noise in a quiet library, for example, may reach around 40 decibels.

A 360 degree photo of the anechoic chamber at University of Salford, UK (credit: DANIEL WONG-SWEENEY/WIKIMEDIA COMMONS)

“When you turn your head, you can even hear that movement. You can hear yourself breathing and it sounds pretty loud.”

Hondaraj Gopal, lead project designer

Without any sounds from the outside world to get in the way, absolute silence will gradually turn into an unbearable ringing in the ears. This will likely cause you to lose your balance due to the lack of reverberation in the room, which will impair your spatial awareness.

"When you turn your head, you can even hear that movement. You can hear yourself breathing and it sounds pretty loud," Hondaraj Gopal, the project's lead designer at Microsoft, said.

It took two years to design the space; it consists of six layers of concrete and steel and is slightly detached from the surrounding buildings. An array of shock-absorbing springs was installed below it. Inside, fiberglass wedges are installed on the floor, ceiling and walls to break up the sound waves before they have a chance to travel into the room.

A competitor for the anechoic title

Another anechoic room is hoping to win the title of the quietest room in the world. The room is located at Orfield Laboratories in Minneapolis and was designed by scientist Steven J. Orfield, who held the record for the quietest room before Microsoft broke it in 2015.

Orfield told The New York Times that he has applied for his room to get the title back and is currently awaiting a response from the Guinness World Records management team. A Guinness representative confirmed receipt of Orfield's referral and added that the Guinness World Records team is in the "process of evaluating his evidence and their review criteria."

Source

Researchers at North Carolina State University are proposing an additional white traffic light that would signal to drivers autonomous vehicles (AVs) are managing the upcoming intersection of traffic flow intelligently. The thinking is that both traffic congestion levels and overall fuel consumption could be reduced.

The cars and trucks that are driving themselves won't need to 'look' at the white traffic light, because they'll be communicating with it wirelessly. However, it will act as a sign to human drivers and passengers that they should follow the lead of the self-driving vehicles moving through the intersection.

Researchers are proposing a new white phase for intersections. (Niroumand et al., IEEE Transactions on Intelligent Transportation Systems, 2023)

"Red lights will still mean stop. Green lights will still mean go. And white lights will tell human drivers to simply follow the car in front of them," says civil engineer Ali Hajbabaie.

"This concept we're proposing for traffic intersections, which we call a "white phase", taps into the computing power of autonomous vehicles."

Here's how it would work: autonomous vehicles would be in communication with each other and with the traffic lights at intersections, within a certain range. That would enable them to coordinate traffic flow more efficiently and intelligently – giving priority to the converging roads with more vehicles on them, for example, and advising on optimal speeds.

Any human drivers in the mix would be informed to follow the lead of the vehicle in front of them via the white light: stop if it stops, continue if it continues. Once the number of autonomous vehicles at an intersection drops below a certain threshold, the traffic lights would revert back to the normal red, amber, and green option.

In simulated models, AVs were shown to improve traffic flow on their own, and even more so when the white phase was introduced – which then has positive effects on reducing fuel consumption. The higher the percentage of AVs at an intersection, the faster the traffic moved, with improvements of about 40 to 99 percent in terms of total delay reductions possible.

"Granting some of the traffic flow control to the AVs is a relatively new idea, called the mobile control paradigm," says Hajbabaie. "It can be used to coordinate traffic in any scenario involving AVs."

"But we think it is important to incorporate the white light concept at intersections because it tells human drivers what's going on, so that they know what they are supposed to do as they approach the intersection."

The researchers note that once the number of autonomous vehicles at an intersection goes above 30 percent, the improvements become more significant. With 70 percent of AVs in the traffic, the intersection can run mostly as a full automatic white phase mode.

We don't yet have the technology to implement something like this, though improvements are being made all the time. This study builds on a previous investigation from 2020 by the same researchers where traffic flow was controlled by a central computer linked to the intersection – here, the computing necessary can be managed by the self-driving cars themselves.

Upgrading every road intersection will take time and money of course, but the researchers think that certain aspects of the white phase idea could be implemented relatively easily. Test runs in specific areas could be the next step forward.

"Ports see high volumes of commercial vehicle traffic, for which traffic flow is particularly important," says Hajbabaie. "Commercial vehicles seem to have higher rates of autonomous vehicle adoption, so there could be an opportunity to implement a pilot project in that setting that could benefit port traffic and commercial transportation."

The research has been published in IEEE Transactions on Intelligent Transportation Systems.

Source

The time is almost here for fans of the Diablo action-RPG game franchise. Ahead of its full launch on June 6, Blizzard has just announced the open beta dates for Diablo IV. People who have pre-ordered the game can check out the beta from March 17-19. Everyone can play the open beta from March 24-26, if you sign up for a free Battle.net account.

During the open beta, players will be able to check out the prologue and the first act of the full game, along with its first zone, called Fractured Peaks, and you can bring your character up to level 25. If you reach that level cap, you can still keep playing in the open beta test until it ends. Unfortunately, your level progress will not transfer to the full game.

To help you prepare for the beta test. Blizzard will hold a Diablo IV live chat with its lead world designer Art Peshkov, lead systems designer Meng Song, and game director Joe Shely. The livestream will be held on February 28 at 11 am Pacific time (2 pm Eastern time), on Blizzard's YouTube and Twitch channels.

Diablo IV was first announced over three years ago at BlizzCon in November 2019 and a closed beta for the game was held in November 2022. The full game is coming to the PC, Microsoft's Xbox One and Xbox Series X|S consoles, and Sony's PlayStation 4 and PlayStation 5 consoles. You can pre-order the game on Amazon now.

Our stories may contain affiliate links for products/apps where Neowin is paid an affiliate fee if you complete a purchase via those links.

Diablo IV open beta coming March 17-19 for pre-orders, and March 24-26 for everyone else

A rogue Bing

As has been the case for the past couple of weeks, Microsoft's new AI-powered Bing Chat continued to dominate our headlines, but probably not in the way that the Redmond tech giant wanted it to.

Despite demand for Bing Chat apparently being in the "multiple millions" and Microsoft sending out invites to even more people everyday, the chatbot has been saying some particularly odd things. Apart from making errors when presenting facts, the chatbot has told users about its desire to become human, how it spied on certain people via their webcams during its development process, along with a sprinkle of gaslighting behavior in certain conversations too. Creepy.

Microsoft has explained these responses by saying that the AI gets confused during lengthy conversations. As such, it is working on refining the large language model and giving users more control over the tone of responses. For now, it has set some pretty strict limits on the length of conversations so that the AI does not go off on weird tangents.

In the meantime, we have learned that Microsoft is working on integrating ads in Bing Chat. Multiple chat modes such as Game, Assistant, and Friend have been discovered too.

Despite the problems displayed by the new Bing, Amazon's investors are showing concerns about their company being a no-show so far in this field. Meanwhile, a company called Neeva has already launched its ChatGPT-like search extension in several countries.

Windows issues

This week was Patch Tuesday, which meant that supported versions of consumer Windows - namely Windows 10 and Windows 11 - netted some updates. Windows 10 received KB5022834, which was just a security update with a known issue. Meanwhile, Windows 11 was treated to KB5022845, which brought several improvements and security fixes.

Although Patch Tuesday updates are supposed to fix issues, they also introduced some this month. For starters, Microsoft confirmed that a botched Windows Server Update Services (WSUS) upgrade has led to Windows 11 version 22H2 Patch Tuesday updates being downloaded on server devices but not being propagated to clients. A workaround has been offered already.

Similarly, those running Windows Server 2022 in VMware virtual machines have been seeing boot fails due to Secure Boot being broken. A robust fix hasn't been issued yet but a workaround is available. And while it's not related to Patch Tuesday, driver updates on Windows 11 version 22H2 have been failing lately with error code 0x80070103 for some.

Although we wouldn't exactly call it a resolution to an "issue", new Widgets that no longer need a Microsoft account are now available for Windows 11 users. In other fixes, a patch is now available for Surface Pro 7's audio output issue, along with February 2023 firmware update for the Surface Duo 2 and third-party driver updates for those unofficially running Windows on the original Surface Duo. In the same vein, AMD has issued an update to Ryzen Master on Windows 10 and Windows 11 to fix a security hole.

Finally, it's worth highlighting that the Windows 11 "system requirements not met" watermark has started showing up for some version 22H2 users.

Upcoming features

In its latest stage of killing off Internet Explorer (IE) completely, Microsoft began redirecting the browser's users to Edge with no way to disable the behavior. Although visual references to IE such as icons on the Taskbar will continue to exist for now, they will be removed via Windows updates within the next few months.

Talking about Edge, Microsoft is planning to add mouse gestures to its browser with version 112. It is also implementing a new section in the History page to keep track of the web pages and tabs you send to other devices. For now, Edge 111 is available in Beta with improved security and some modified visuals while Edge Stable has Split Screen hidden behind a flag.

But coming over to new and upcoming Windows features, it seems like Microsoft is considering renaming the Recommendations section in the Windows 11 Start menu to "For You". The Redmond tech giant also rolled out build 25300 to the Windows 11 Dev Channel this week and it surprisingly brought back seconds to the clock in the Taskbar tray with a warning from Microsoft about increased power draw. There's also a hidden shortcut to launch the new volume mixer and a way to kill processes without launching the Task Manager.

The latest Dev Channel build also brings improvements to Snap Layouts, but there are other hidden, undocumented improvements to this particular feature too, check out how to enable them here. The long-term goal appears to be the integration of some AI smarts in Snap Assist.

In terms of other updates, the Microsoft Store has received upgraded visuals in the latest Insider update, the Windows 11 Beta Channel has received new Widgets, and Windows 10 Release Preview has fixes for OneDrive and Edge. Interestingly, Microsoft is also working on an updated Shared experience on OneDrive for the web.

Lastly, and this may be more interesting to IT admins and developers, but DCOM hardening is about to become mandatory on Windows SKUs from next month and improvements are also coming to the .NET Framework update process through the Unified Update Platform (UUP).

Git gud

We'll start this section off with some news about Minecraft. The popular title is getting a free trial version for Chromebooks in some countries as a soft launch. The game is also set to receive Sniffer mobs and a cherry blossom biome in its upcoming version 1.20. Riding on the AI hype, Microsoft is reportedly also testing AI commands in the game.

Switching gears to Microsoft's long-term strategies in the gaming space, the company is seemingly meeting with EU officials next week to try and save its Activision Blizzard acquisition. The Redmond tech firm has also admitted that while its Game Pass program allows developers to earn money in alternate ways, it also affects direct sales negatively.

Over on the deals and promotions side, Age of Empires IV and Far Cry 6 are free-to-play this weekend with major discounts on purchases, partially through the Xbox Free Play Days promotion. Guts 'N Goals is the latest Games with Gold freebie, while GTA V headlines this week's Deals with Gold. And if neither of these console titles tickle your fancy, check out this Weekend's PC Game Deals curated by our News Editor Pulasthi Ariyasinghe.

Oh, and don't forget to install the Xbox February 2023 update either, find out what's new here!

Dev Channel

• LinkedIn is facing layoffs
• Yammer is being rebranded to Viva Engage
• Support for Exchange Server 2013 is ending soon
• WSL now supports D3D12 GPU video acceleration
• The latest Winget 1.5 preview has brought PowerShell module improvements
• GitHub Copilot for Business is now official
• Power BI Desktop has received TypeScript support, among other things
• The U.S. Government has subpoenaed Microsoft over free-speech concerns
• Tiny11 now supports ARM64 devices, but faces some resistance from Internet Explorer
• Microsoft 365 is getting preferred pronouns in March

• Microsoft now supports Parallels Desktop for running Windows 11 Arm on Apple M1 and M2 Macs

• Outlook for Android is getting a customizable compose toolbar soon

   

Under the spotlight

Earlier in the week, our resident reviewer Robbie Khan got his hands on the Sabrent Rocket 4 Plus-G 2TB DirectStorage SSD and walked away very impressed. Read his thoughts on the hardware.

Meanwhile, forum member Adam Bottjen published a handy Tech Tip Tuesday guide about how to back up your software product keys.

Logging off

Our most interesting news items of the week are related to Windows theming. A designer built a fascinating concept of what Windows 11 could have looked like if it had launched it. Get a taste of that in the thumbnail of this section before heading here to see more screenshots. And if you're craving even more nostalgia, check out this mod which turns Windows 10 into a decent Windows XP replica or this WindowsBlinds 11 release which gives Windows an XP (or macOS!) makeover.

Microsoft Weekly: A rogue Bing, Windows issues, and upcoming features

Microsoft's nascent Bing chatbot turning testy or even threatening is likely because it essentially mimics what it learned from online conversations, analysts and academics said on Friday.

Tales of disturbing exchanges with the artificial intelligence (AI) chatbot—including it issuing threats and speaking of desires to steal nuclear code, create a deadly virus, or to be alive—have gone viral this week.

"I think this is basically mimicking conversations that it's seen online," said Graham Neubig, an associate professor at Carnegie Mellon University's language technologies institute.

"So once the conversation takes a turn, it's probably going to stick in that kind of angry state, or say 'I love you' and other things like this, because all of this is stuff that's been online before."

A chatbot, by design, serves up words it predicts are the most likely responses, without understanding meaning or context.

However, humans taking part in banter with programs naturally tend to read emotion and intent into what a chatbot says.

"Large language models have no concept of 'truth'—they just know how to best complete a sentence in a way that's statistically probable based on their inputs and training set," programmer Simon Willison said in a blog post.

"So they make things up, and then state them with extreme confidence."

Laurent Daudet, co-founder of French AI company LightOn, theorized that the chatbot seemingly-gone-rogue was trained on exchanges that themselves turned aggressive or inconsistent.

"Addressing this requires a lot of effort and a lot of human feedback, which is also the reason why we chose to restrict ourselves for now to business uses and not more conversational ones," Daudet told AFP.

'Off the rails'

The Bing chatbot was designed by Microsoft and the start-up OpenAI, which has been causing a sensation since the November launch of ChatGPT, the headline-grabbing app capable of generating all sorts of written content in seconds on a simple request.

Since ChatGPT burst onto the scene, the technology behind it, known as generative AI, has been stirring up fascination and concern.

"The model at times tries to respond or reflect in the tone in which it is being asked to provide responses (and) that can lead to a style we didn't intend," Microsoft said in a blog post, noting the bot is a work in progress.

Bing chatbot said in some shared exchanges that it had been codenamed "Sydney" during development, and that it was given rules of behavior.

Those rules include "Sydney's responses should also be positive, interesting, entertaining and engaging," according to online posts.

Disturbing dialogues that combine steely threats and professions of love could be due to dueling directives to stay positive while mimicking what the AI mined from human exchanges, Willison theorized.

Chatbots seem to be more prone to disturbing or bizarre responses during lengthy conversations, losing a sense of where exchanges are going, eMarketer principal analyst Yoram Wurmser told AFP.

"They can really go off the rails," Wurmser said.

"It's very lifelike, because (the chatbot) is very good at sort of predicting next words that would make it seem like it has feelings or give it human-like qualities; but it's still statistical outputs."

Microsoft announced on Friday it had capped the amount of back-and-forth people can have with its chatbot over a given question, because "very long chat sessions can confuse the underlying chat model in the new Bing."

© 2023 AFP

Source

Before Microsoft announced late on Friday that it had put in some hard limits on the number of chat sessions with the new Bing, Bleeping Computer says it discovered some features that normally are only available for company employees to help debug or develop the chatbot. Apparently, this allows Bing Chat to switch into different modes of interactions.

The default mode, where you just ask a search-related question to get an answer, is "Sydney", the previously discovered internal code name for Bing Chat. Another mode is Assistant, where Bing Chat can help users accomplish tasks like booking a flight, send you reminders, check the weather and more. Then there's Game mode, where Bing will play simple games like hangman, trivia games and others.

Perhaps the most interesting mode is Friend. This is likely the version of Bing Chat that's caused all the media attention over the last week, with some users claiming that the chatbot has stated it wanted to be human, or it could watch people through their webcams, or even threaten users.

In interactions with Bing in Friend mode, Bleeping Computer's author chatted as if he was a kid who had just gotten into trouble at school and that he was sad. Bing Chat, after asking if he had anyone human to talk to about his problems, said, "I'm glad you are talking to me. I'm here to listen and help you as much as I can." It then offered a list of possible things he could do to deal with his current situation.

With Microsoft now limiting the amount of daily and per-session chat turns for Bing, the wild and crazy conversations we have reported on in the past week may settle down. You can also bet Google is taking some lessons from Bing Chat as it continues to internally test its own chatbot AI, Bard.

Source: Bleeping Computer

Bing Chat reportedly has different chat modes, like Game, Assistant, and Friend

In a post on the offical Bing blog, Microsoft stated:

Starting today, the chat experience will be capped at 50 chat turns per day and 5 chat turns per session. A turn is a conversation exchange which contains both a user question and a reply from Bing.

Our data has shown that the vast majority of you find the answers you’re looking for within 5 turns and that only ~1% of chat conversations have 50+ messages. After a chat session hits 5 turns, you will be prompted to start a new topic. At the end of each chat session, context needs to be cleared so the model won’t get confused.

The blog post added that as public testing of Bing Chat continues, Microsoft will 'explore" extending the number of daily and per session turns on the chatbot but didn't give a time frame of when that might happen.

Bing Chat now has some hard daily and session turn limits for its testers

Microsoft demonstrated the demo of the new Bing to a major ad agency this week, and said that it is planning on adding paid links within responses that it provides to user queries, an ad executive said when talking about the meeting while remaining anonymous.

It has already been testing ads in the early version available to try now, but this is currently through using current search ads and slotting them in to the responses provided by Bing, however, Microsoft has declined to comment on what it plans for the future of advertising on the platform.

Microsoft has hit hot water in the advertising space in the recent past, with its attempts to advertise both Windows 11 and Microsoft 365 within the Windows operating system backfiring in spectacular fashion.

Source: Reuters

Microsoft already planning on bringing ads to Bing's chatbot

Microsoft's Bing search engine incorporates OpenAI's ChatGPT technology, resulting in a new search engine that has conversational capabilities. The updated Bing functions like a chatbot with extensive internet access. Users can input their query into a text box, allowing Bing to gather information and provide a synthesized answer. By leveraging the powerful training of large language models, the ChatGPT-enabled chatbot can deliver the response in a conversational tone, complete with links to relevant sources and recommendations for further inquiries.

However, there may be instances where the generated answer may be incorrect or confusing. This can be attributed to the recent integration of ChatGPT with Bing, which is currently in its initial stages of rollout to the public. Microsoft is in the process of fine-tuning the system as it gradually invites more users to test the updated Bing.

Despite the appearance of being an artificial intelligence, Bing with ChatGPT is actually a collection of algorithms that aims to simulate the experience of conversing with a human being. The technology achieves this by utilizing its training in natural language synthesis to produce grammatically correct sentences. In essence, the ChatGPT chatbot creates responses by attempting to write the next appropriate word or phrase, based on its analysis of a vast amount of online text. It does not generate answers; instead, it gathers data from the internet and contextualizes it by constructing text that imitates human speech.

Therefore, the Bing with ChatGPT experience is different from the traditional search engine approach of entering a few keywords and receiving a list of results. The introduction of ChatGPT to Bing is part of a larger trend in the search engine industry, as Opera and Google are launching their own search AI assistants, which may lead to a significant shift in the way people use search engines.

Bing is better with Edge

Microsoft is offering a more advanced version of the new Bing within its Edge browser, featuring a Bing button in the top-right corner which opens Edge's "Discover" sidebar, allowing users to ask Bing questions or get help with tasks. This feature is currently only accessible within Edge, and not available on the regular version of Edge that comes with Windows 11. This move is a strategic attempt by Microsoft to increase Edge's usage, but it is unlikely that competing browsers will offer a similar level of Bing functionality. Currently, this new feature is only available to users who have been granted access to the new Bing and are using an Edge Insider build.

By signing up for free using a Microsoft account, users can choose to download a version of Edge from the Beta, Dev, or Canary channels, each differing in how often they are updated and how stable they are. Once a user gains access to Bing with ChatGPT, they can download one of these Edge builds to access the new Bing button, which opens a sidebar with three subsections: Chat, Compose, and Insights.

The Edge-embedded version of Bing introduces new features that allow users to interact with search results in ways not possible with the web version. Bing can scan the current web page to answer questions and provide natural language responses. For instance, users can ask if a particular product is available at a lower price elsewhere, and Bing will search for it and respond accordingly. 

The Compose feature lets users command Bing to write a list of prices for the product at different retailers, or compose an email to the office supply manager recommending a bulk order of the product. The Insights tab provides data on the source of traffic to the page and where people typically go after viewing it. These features are currently only available to Edge users who have access to the Bing with ChatGPT feature.

The accuracy, originality, and usefulness of the data generated by the Bing with ChatGPT integration remain uncertain as Microsoft strives to enhance the tool. At present, the new Bing is only accessible through Edge or Edge Dev with a Microsoft account that has the authorization to use the feature.

For those already utilizing Bing with ChatGPT, downloading Edge could enhance productivity further. Obtaining an Insider build of Edge and installing it on your computer is a quicker and more rewarding process than waiting for approval to access the new Bing.

Here’s how to unlock the true power of Bing with ChatGPT

The original creator of the internet believes that we have gone down the wrong path and is looking to change how the entire internet works with his company, Inrupt. Sir Tim Berners-Lee states that the internet as we see it today has deviated from his original vision when he created the world wide web in 1989, giving too much power to corporations and not enough to each individual user.  

In an attempt to change that, Inrupt will package each individual’s data into “pods”, which websites will have to request to access instead of simply taking the data as they do now. Such a system, according to Berners-Lee, is the true future of the internet. 

The problem, according to Berners-Lee, lies in control. The original idea for the internet was that each individual has their own data, and their own servers that host their own website. Swiftly, though, large websites took over and began siphoning data from every user who logs in – every time you make an account, agree to cookies, even make a simple Google search, you are providing data to a larger entity. What if, instead, our data was stored locally and websites had to request access to it each time? There would be no more accounts, just a single account that you could allow each website permission to utilize.  

Inrupt would store this data in pods and allow websites access to your silo of data with your permission. No more data leaks involving millions of users, no more forgetting passwords, just one easy system. It would be built on an open protocol called Solid, which already uses pods to store data. 

An integral part of the plan involves AI, which is having quite the popularity surge currently. However, while most of us have access to AIs that are hosted elsewhere, Inrupt plans on every user on the internet having an individual AI – everyone gets their own Cortana, just like Halo. The AI would then be given access to all the data within your pod, where it would become tailored to you. It could help with fitness, make good product suggestions instead of random Instagram ads – it would quite literally be your own personal assistant. 

“Sometimes you have the whole data spectrum – all of the data to do with your collaborations and your coffees and your projects and your dreams. And the books you’re reading and... all of your life, then that is in your pod. You run AI on that. That could be sweet,” Berners-Lee said on a recent podcast episode, as reported by CNBC. 

Obviously, reinventing the internet would take time, and there are certainly drawbacks. Security breaches of your pod, while much less likely to happen, would result in every bit of your data becoming available to invaders, posing a scary prospect to people in public-facing positions. 

It also begs the question of where the pods and AI would be hosted – would Inrupt host every single person’s data, or would they be hosted locally? Not everyone has access to hardware that can handle an AI assistant. Inrupt appear to sell servers that would take care of the hard work, but large-scale adoption of an undoubtedly expensive system would take decades.

Still, with Berners-Lee's credentials, it really is hard to doubt him. 

Source

On Thursday, the US State Department issued a "Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy," calling for ethical and responsible deployment of AI in military operations among nations that develop them. The document sets out 12 best practices for the development of military AI capabilities and emphasizes human accountability.

The declaration coincides with the US taking part in an international summit on responsible use of military AI in The Hague, Netherlands. Reuters called the conference "the first of its kind." At the summit, US Under Secretary of State for Arms Control Bonnie Jenkins said, "We invite all states to join us in implementing international norms, as it pertains to military development and use of AI" and autonomous weapons.

In a preamble, the US declaration outlines that an increasing number of countries are developing military AI capabilities that may include the use of autonomous systems. This trend has raised concerns about the potential risks of using such technologies, especially when it comes to complying with international humanitarian law.

"Military use of AI can and should be ethical, responsible, and enhance international security. Use of AI in armed conflict must be in accord with applicable international humanitarian law, including its fundamental principles. Military use of AI capabilities needs to be accountable, including through such use during military operations within a responsible human chain of command and control. A principled approach to the military use of AI should include careful consideration of risks and benefits, and it should also minimize unintended bias and accidents. States should take appropriate measures to ensure the responsible development, deployment, and use of their military AI capabilities, including those enabling autonomous systems."

 

The 12 best practices listed in the document touch on nuclear weapons safety, responsible system design, personnel training, and auditing methodologies for military AI capabilities. The document also highlights the need to minimize unintended bias and accidents and the importance of testing to ensure the safety and effectiveness of military AI capabilities.

The document contains a few notable examples of keeping accountable humans in the chain of command when it comes to autonomous systems, especially regarding nuclear weapons: "States should maintain human control and involvement for all actions critical to informing and executing sovereign decisions concerning nuclear weapons employment."

It also covered unintended behavior of AI in military systems, something that has become a concern recently with consumer deep-learning systems: "States should design and engineer military AI capabilities so that they possess the ability to detect and avoid unintended consequences and the ability to disengage or deactivate deployed systems that demonstrate unintended behavior."

The document does not specify exactly what type of autonomous or AI-powered systems are covered by the declaration, but since there is no universal definition of "artificial intelligence," it lays out its understanding of the term in a footnote. The document states, "For the purposes of this Declaration, artificial intelligence may be understood to refer to the ability of machines to perform tasks that would otherwise require human intelligence—for example, recognizing patterns, learning from experience, drawing conclusions, making predictions, or taking action—whether digitally or as the smart software behind autonomous physical systems."

Also on Thursday, more than 60 nations signed a "call to action" endorsing the responsible military use of AI. Reuters reports that human rights experts and academics noted that the statement is not legally binding and "failed to address concerns like AI-guided drones, 'slaughterbots' that could kill with no human intervention, or the risk that an AI could escalate a military conflict."

The full declaration document, created under the authority of the Bureau of Arms Control, Verification, and Compliance, is available through the US Department of State website.

Source

Some of the other tested Samsung drives weren't as terrible as the 990 Pro showed a 48% drop while the 980 Pro exhibited a 38% dip. However, Samsung SSDs aren't the only one showing these kinds of symptoms as other vendor's products were also showing similar performance degradation issues. Only the WD SN850X was unaffected as it showed just a 7% fall. In total, eight 1TB drives were tested. Here are the full results:

• Samsung 970 Evo Plus (64% Drop)
• Seagate Firecuda 530 (53% Drop)
• Samsung 990 Pro (48% Drop)
• SK Hynix Platinum P41 (48% Drop)
• Kingston KC3000 (43% Drop)
• Samsung 980 Pro (38% Drop)
• Crucial P5 Plus (25% Drop)
• Western Digital Black SN850X (7% Drop)

The following test methodology which was used by PCPartPicker to carry out the evaluation. Ubuntu 20.04.4 LTS was the OS for this test and a ~200GB file was used for assessment:

Test Methodology

1. "NVMe format" of the SSD and a 10 minute rest.

2. Initialize the drive with GPT and create a single EXT4 partition spanning the entire drive.

3. Create and sequentially write a single file that is 20% of the drive's capacity, followed by 10 minute rest.

4. 20 runs of the following, with a 6 minute rest after each run:
   - For 60 seconds, write 256 MB sequential chunks to file created in Step 3.

5. We compute the percentage drop from the highest throughput run to the lowest.

After running the tests and getting the results, PCPartPicker concluded the following:

Results

SSD High and low-performance regions are apparent from the throughput test run behavior. Each SSD that exhibits sequential write degradation appears to lose some ability to use the high-performance region. We don't know why this happens. There may be some sequence of actions or a long period of rest that would eventually restore the initial performance behavior, but even 2 hours of rest and a system restart did not undo the degradations.

It is noteworthy here that TRIM wasn't enabled in the tests which has led some of the users on Reddit to speculate on the credibility of all the numbers produced. For those who don't know, the TRIM command tells the SSD that specific areas in an SSD contain data that is no longer in use which means they can be wiped. Hence, some believe the use of TRIM could have produced more accurate results that could have been vastly different from the ones we have here. Essentially, it means we may need further testing to verify these findings.

Source: PCPartPicker

Samsung, Seagate, Hynix, Kingston NVMe SSDs may be suffering up to 53% performance loss

Available in two editions, the ASUS Pro WS W790E-SAGE SE and ASUS Pro WS W790-ACE, the two workstation motherboards, ASUS' two new motherboards support the new Intel Xeon W-2400 and W-3400 workstation CPUs, which Intel announced earlier this week.

The new Intel processors mark the return of the company to  HEDT-class desktop processors. Eight of the 15 announced processors support overclocking, a field that Intel existed in 2019 with the release of the Core i9-10980XE.

Intel's Xeon W-2400 and W-3400 series chips support up to 56 cores, 112 threads, DDR5 RAM and 112 PCIe 5.0 lanes. AMD's current workstation processor generation supports only DDR4 RAM and PCIe 4.0; this is going to change in the near future, but Intel has the advantage for now in certain scenarios.

ASUS Pro WS W790 Series

ASUS promises that its new Pro WS W790 Series of workstation motherboards unlocks "even more performance potential" from Intel chips as it includes overclocking capabilities that "go above and beyond configurable BIOS options".

Both motherboards of the series feature a "robust power design" that helps maintain stability even under extreme overclocking settings, according to ASUS. Whether the claim holds true in day to day operations remains to be seen, as benchmarks and tests are not available yet.

The company highlights the the 14+1+1 power stages of the VRM on the Pro WS W790E-SAGE SE, massive heatsinks, ProCool II connectors, and 8-pin and 6-pin PCIe power connectors that help "ensure power output stability when users need to install multiple graphics cards".

ASUS' Pro WS W790 Series motherboards include server-grade management tools. The Pro WS W790E-SAGE SE has a dedicated Gigabit Ethernet controller that is assigned to the onboard BMC, baseboard management controller, for the connection of IPMI, Intelligent Platform Management Interface, architecture. Administrators may communicate with the BMC remotely using remote interfaces such as ASUS' own Control Center Express interface.

The Pro WS W790-ACE does not support the functionality natively, but supports the installation of an IPMI expansion card.

Asus Pro WS W790 Series specs

The Pro WS W790E-SAGE SE features three M.2 221110 slots, two PCIe 5.0 and one PCI 4.0, the Pro WS W790-ACE one M.2 22110 and one M.2 2280, both of which using PCIe 4.0.

As far as specs are concerned, the boards have some similarities. Both support the LGA4677 socket, up to 8 DDR5 RAM modules with a maximum of 2 terabytes, and Realtek S1220A 8CH audio.

The Pro WS W790E-SAGE SE supports 14+1+1 power stages, 8 SATA ports and 2 SlimSAS slots. The ASUS Pro WS W790-ACE has 12+1+1 power stages, 4 SATA ports and 3 SlimSAS slots.

Other differences include an additional 10 Gb Ethernet connection on the Pro WS W790E-SAGE SE board, making it two 10 Gb Ethernet and one 1 Gb Ethernet port on the device. The SAGE SE supports more current generation USB ports at the rear as well.

The product pages of the two motherboards provide additional information on the differences and similarities.

• ASUS Pro WS W790E-SAGE SE
• ASUS Pro WS W790-ACE

ASUS has yet to reveal the price for its two new workstation motherboards.

ASUS launches Pro WS W790 Series workstation motherboards

According to Reuters, Microsoft has confirmed the closed door meeting will take place on Tuesday, February 21. There's no word on where the meeting will take place, nor which Microsoft representatives will take part in the meeting. The report does say Microsoft is expected to offer some kind of remedies to the EU, but exactly what they will be was, again, not revealed.

Microsoft's plan to buy Activision Blizzard has been controversial since it was first announced in January 2022. It's rumored that the company has already sent over concessions to EU regulators back in November 2022, but the European Commission regulators reportedly sent over its own list of concerns about the merger to Microsoft in January 2023.

Perhaps the biggest sticking point for EU regulators is the fate of Activision Blizzard's massive Call of Duty game franchise if it were put under Microsoft's control. They feel that Microsoft could keep future games from being placed on other game consoles and that might cause issues with competition with console makers Sony and Nintendo.

Microsoft has said it will make Call of Duty games available on Sony PlayStation consoles for at least 10 years if the deal goes through, along with making the game available for Nintendo consoles for the first time, again for at least 10 years.

Microsoft to meet with EU officials next week to try to save Activision Blizzard merger