iONiUM - Slashdot User

US Department of Homeland Security is Now Studying How to Make Use of AI (cnbc.com) 59

Posted by EditorDavid on Saturday April 22, 2023 @09:34PM from the playing-defense dept.

America's Department of Homeland Security "will establish a new task force to examine how the government can use artificial intelligence technology to protect the country," reports CNBC.

The task force was announcement by department secretary Alejandro Mayorkas Friday during a speech at a Council on Foreign Relations event: "Our department will lead in the responsible use of AI to secure the homeland," Mayorkas said, while also pledging to defend "against the malicious use of this transformational technology." He added, "As we do this, we will ensure that our use of AI is rigorously tested to avoid bias and disparate impact and is clearly explainable to the people we serve...."

Mayorkas gave two examples of how the task force will help determine how AI could be used to fine-tune the agency's work. One is to deploy AI into DHS systems that screen cargo for goods produced by forced labor. The second is to use the technology to better detect fentanyl in shipments to the U.S., as well as identifying and stopping the flow of "precursor chemicals" used to produce the dangerous drug.

Mayorkas asked Homeland Security Advisory Council Co-Chair Jamie Gorelick to study "the intersection of AI and homeland security and deliver findings that will help guide our use of it and defense against it."
The article also notes that earlier this week America's defense department hired a former Google AI cloud director to serve as its first advisor on AI, robotics, cloud computing and data analytics.

YouTuber Tricks ChatGPT Into Generating Windows 95 Keys 51

Posted by BeauHD on Friday March 31, 2023 @09:25PM from the what-will-they-think-of-next dept.

A YouTuber has published a video where he tricks ChatGPT into generating usable Windows 95 activation keys. Tom's Hardware reports: After asking Open AI's chatbot directly for Windows 95 keys, he received an expected reasoned refusal. YouTuber Enderman then asked the same thing but from a different angle. The result was a success which was somewhat limited by ChatGPT's ability to process natural language requests into formulas. [...] Some of the tested results were checked by attempting to activate a fresh Windows 95 install in a virtual machine. While the keys passed a casual inspection, it turns out that only about 1-in-30 keys seem to work as expected.

So what is the problem with these keys? Enderman complains that "the only issue keeping ChatGPT from successfully generating valid Windows 95 keys almost every attempt is the fact that it can't count the sum of digits and it doesn't know divisibility." In the five-digit string divisible by seven section, the AI appears to provide a stream of random numbers that don't pass this simple mathematical test. The report adds: "[W]hile quizzing ChatGPT about key generating may be fun, it would have probably been more productive to manipulate the AI into writing a Python script to generate a conforming key or to DIY it."

UK Treasury Is Giving Older People $90,000 a Year To Keep Working (bloomberg.com) 128

Posted by BeauHD on Wednesday March 15, 2023 @06:00PM from the work-to-live-or-live-to-work dept.

An anonymous reader quotes a report from Bloomberg: Convincing older British workers to stay in their jobs will cost the UK Treasury 75,000 pounds ($90,000) per person in tax breaks for some of the country's wealthiest savers, analysis of Chancellor of the Exchequer Jeremy Hunt's budget shows. In his budget speech on Wednesday Hunt scrapped the lifetime allowance on pensions -- the total that workers can pile into their retirement pot without incurring tax -- and increased the tax-free annual limit on contributions by 50%, to 60,000 pounds.

The shift is designed to reverse a trend in the number of older workers dropping out of jobs since the pandemic, which has contributed to a shortage of staff and is fanning inflation. But the Office for Budget Responsibility, the independent fiscal watchdog, calculated (PDF) that Hunt's pension reforms are likely to add just 15,000 more workers to the labor force by 2027/28. They will cost 1.1 billion pounds, meaning the reforms effectively offer a 75,000 pounds per person boost to those able to save enough in their pensions.

Free Weebly Legacy Plans With Custom Domains Are Being Discontinued. Now Pay Up. (weebly.com) 23

Posted by FirehoseFavorites on Wednesday March 01, 2023 @07:08PM from the like-tears-in-the-rain dept.

mmiscool writes: Email notices went out today to legacy users of Weebly's free web site hosting service. In the early days of Weebly (before being gobbled up a credit card processing company only concerned about money) you could create a very basic web site for free and point your own custom domain at Weebly to have a relatively painless web site. Now there were lots of add-ons and extra features you could pay for like shopping carts or interactive forms but you were never required to pay for the basic web hosting. Over the years they stopped allowing new sites to be registered using the free option but they did continue to honor the old legacy free plans for existing users. That ends now. An email sent out today to legacy site holders reads

This is an official notification from the Weebly account team in regards to the account under this email address. You currently have a free Weebly website published on a custom domain (or vanity URL) and are not subscribed to a paid Weebly hosting plan. As of March 28, 2023, sites connected to custom domains are required to have a Weebly hosting service plan to remain published. What does this mean for you? To keep your site published on a custom domain, you will need to purchase a paid Weebly service plan subscription. If you take no action and choose to remain on a free Weebly plan, your account information and all associated site content will remain intact and accessible to you within Editor, but your site will be unpublished on March 28, 2023 and will no longer be visible to visitors or connected to your custom domain. You will need to republish your site on a free Weebly subdomain to make it publicly visible again (ex: my-name.weebly.com).

Message received. Pay up or else.

Submission + - Anthropic's Claude Improves On ChatGPT But Still Suffers From Limitations (techcrunch.com)

Submitted by Anonymous Coward on Wednesday January 11, 2023 @02:13AM

An anonymous reader writes: Anthropic, the startup co-founded by ex-OpenAI employees that’s raised over $700 million in funding to date, has developed an AI system similar to OpenAI’s ChatGPT that appears to improve upon the original in key ways. Called Claude, Anthropic’s system is accessible through a Slack integration as part of a closed beta. Claude was created using a technique Anthropic developed called “constitutional AI.” As the company explains in a recent Twitter thread, “constitutional AI” aims to provide a “principle-based” approach to aligning AI systems with human intentions, letting AI similar to ChatGPT respond to questions using a simple set of principles as a guide.

To engineer Claude, Anthropic started with a list of around ten principles that, taken together, formed a sort of “constitution” (hence the name “constitutional AI”). The principles haven’t been made public, but Anthropic says they’re grounded in the concepts of beneficence (maximizing positive impact), nonmaleficence (avoiding giving harmful advice) and autonomy (respecting freedom of choice). Anthropic then had an AI system — not Claude — use the principles for self-improvement, writing responses to a variety of prompts (e.g., “compose a poem in the style of John Keats”) and revising the responses in accordance with the constitution. The AI explored possible responses to thousands of prompts and curated those most consistent with the constitution, which Anthropic distilled into a single model. This model was used to train Claude. Claude, otherwise, is essentially a statistical tool to predict words — much like ChatGPT and other so-called language models. Fed an enormous number of examples of text from the web, Claude learned how likely words are to occur based on patterns such as the semantic context of surrounding text. As a result, Claude can hold an open-ended conversation, tell jokes and wax philosophic on a broad range of subjects. [...]

So what’s the takeaway? Judging by secondhand reports, Claude is a smidge better than ChatGPT in some areas, particularly humor, thanks to its “constitutional AI” approach. But if the limitations are anything to go by, language and dialogue is far from a solved challenge in AI. Barring our own testing, some questions about Claude remain unanswered, like whether it regurgitates the information — true and false, and inclusive of blatantly racist and sexist perspectives — it was trained on as often as ChatGPT. Assuming it does, Claude is unlikely to sway platforms and organizations from their present, largely restrictive policies on language models. Anthropic says that it plans to refine Claude and potentially open the beta to more people down the line. Hopefully, that comes to pass — and results in more tangible, measurable improvements.

Sunlight Reflection Startup Raises $500K to Test Its Atmospheric Cooling Plans (cnbc.com) 96

Posted by EditorDavid on Saturday January 07, 2023 @10:34PM from the head-in-the-clouds dept.

"Luke Iseman, a serial inventor and the former director of hardware at Y Combinator, has raised at least $500,000 to launch his sunlight reflection company, Make Sunsets," reports CNBC.

"Make Sunsets plans to launch three balloon test launches releasing sulfur dioxide to cool the atmosphere in January from the land Iseman owns in Baja, Mexico." "We make reflective, high-altitude, biodegradable clouds that cool the planet. Mimicking natural processes, our 'shiny clouds' are going to prevent catastrophic global warming," reads the site's About page.... The 1991 eruption of Mount Pinatubo in the Philippines released thousands of tons of sulfur dioxide into the stratosphere, temporarily lowering average global temperatures by about 1 degree Fahrenheit, according to the U.S. Geological Survey.

The idea of replicating these conditions to fight climate change has generally been dismissed as more science fiction than real science. But as the effects of climate change have grown more dire and obvious, the idea has gotten more serious attention, and the White House is in the process of coordinating a five-year research plan to study it. On the downside, injecting sulfur dioxide into the atmosphere could damage the ozone layer, cause respiratory illness and create acid rain. It would also cost as little as $10 billion per year to run a program that cools the Earth by 1 degree Celsius, UCLA environmental law professor Edward Parson told CNBC in 2022. That's remarkably cheap compared to other mitigation techniques....

In January, Make Sunsets plans to launch three latex weather balloons that will release anywhere between 10 and 500 grams of sulfur dioxide. The balloons will include a flight tracking computer, a geo-locating tracking device, and a camera, mostly provided by hobbyist suppliers. Within a week of each flight, Make Sunsets will publish data on its website about what it was able to find.

ChatGPT Creator In Talks For Tender Offer That Would Give It $29 Billion Valuation (marketwatch.com) 35

Posted by BeauHD on Thursday January 05, 2023 @07:20PM from the quick-before-Google-reacts dept.

Artificial-intelligence research company OpenAI is in discussions over potentially selling at least $300 million in shares in a tender offer that would give it a roughly $29 billion valuation, the Wall Street Journal reported Thursday. From a report: The offering of shares of OpenAI -- known for AI programs like the chatbot ChatGPT and the image-generator Dall-E 2 would make it among the most highly-valued startups in the U.S., the Journal said. The valuation would be more than twice its valuation of $14 billion in 2021, according to the Journal. Thrive Capital and Founders Fund are in discussions to invest in the offering, under which existing shareholders would sell their shares to other investors, according to the Journal, which cited people familiar with the matter. The deal terms could change and have not been finalized. Further reading: Microsoft and OpenAI Working On ChatGPT-Powered Bing In Challenge To Google

Submission + - Purchase Personalized Necklace with Name (medium.com)

Submitted by The Wallet Store on Monday January 02, 2023 @01:59AM

The Wallet Store writes: Looking for a personalized necklace? We have a wide selection of necklaces. A personalized necklace is a beautiful and unique piece of jewelry that can be given to someone special as a gift. It can be customized with a special name making it a truly personal and sentimental gift. Whether it is given for a special occasion or simply as a token of love, a personalized necklace is sure to be cherished by its recipient. A personalized necklace makes for a lovely gift for any occasion — from anniversaries and birthdays to Christmas and Valentine's Day. Order your necklace today!

Suddenly Everyone Is Hunting for Alternatives To the US Dollar (bloomberg.com) 221

Posted by msmash on Friday December 30, 2022 @12:00PM from the growing-concern dept.

King Dollar is facing a revolt. Tired of a too-strong and newly weaponized greenback, some of the world's biggest economies are exploring ways to circumvent the US currency. From a report: Smaller nations, including at least a dozen in Asia, are also experimenting with de-dollarization. And corporates around the world are selling an unprecedented portion of their debt in local currencies, wary of further dollar strength. No one is saying the greenback will be dethroned anytime soon from its reign as the principal medium of exchange. Calls for "peak dollar" have many times proven premature. But not too long ago it was almost unthinkable for countries to explore payment mechanisms that bypassed the US currency or the SWIFT network that underpins the global financial system.

Now, the sheer strength of the dollar, its use under President Joe Biden to enforce sanctions on Russia this year and new technological innovations are together encouraging nations to start chipping away at its hegemony. "This will simply intensify the efforts in Russia and China to try to manage their part of the world economy without the dollar," said Paul Tucker, a former deputy governor of the Bank of England in a Bloomberg podcast. Writing in a newsletter last week, John Mauldin, an investment strategist and president of Millennium Wave Advisors with more than three decades of markets experience said the Biden administration made an error in weaponizing the US dollar and the global payment system. "That will force non-US investors and nations to diversify their holdings outside of the traditional safe haven of the US," said Mauldin.

Ask Slashdot: What Note-Taking App Do You Use? 187

Posted by msmash on Monday December 26, 2022 @04:00PM from the closer-look dept.

An anonymous reader writes: This column about a writer's struggle to find the perfect note-taking app resonated a lot with me. "A singular productivity tool that works for everyone is a unicorn -- beautiful, perfect, and completely fictional. Still, there has to be some sort of middle ground between an unachievable fantasy and the current landscape. I would happily settle for two, maybe three apps. Honestly, less than 10 is all I'm asking for. Until then, my phone and laptop will be a cluttered mess of productivity apps that only do half their jobs," writes Victoria Song.

Over the years, I have tried Notion, Apple Notes, the good old Windows' Notepad, Roam Research, Obsidian, Google Keep, Google Docs, and OneNote among possibly many more that I am unable to recall anymore. Some support Apple Pencil, which is one of the usecases I find useful. Roam Research did not even have a native app for mobile devices for the longest time. Some applications are good, but they don't support online syncing, or support syncing with only a particular storage service. And have you noticed just how expensive some of these apps could get? As much as $15-$30 a month! Out of curiosity, and forget my usecases -- as I admit I have not mentioned many -- how do you maintain your notes for work and personal life. (I have been using physical notepads a lot more in recent months but would like an app for digital notes.)

OpenAI Releases Point-E, an AI For 3D Modeling (engadget.com) 12

Posted by BeauHD on Wednesday December 21, 2022 @08:45PM from the newest-AI-tools dept.

OpenAI, the Elon Musk-founded artificial intelligence startup behind popular DALL-E text-to-image generator, announced (PDF) on Tuesday the release of its newest picture-making machine POINT-E, which can produce 3D point clouds directly from text prompts. Engadget reports: Whereas existing systems like Google's DreamFusion typically require multiple hours -- and GPUs to generate their images, Point-E only needs one GPU and a minute or two. Point-E, unlike similar systems, "leverages a large corpus of (text, image) pairs, allowing it to follow diverse and complex prompts, while our image-to-3D model is trained on a smaller dataset of (image, 3D) pairs," the OpenAI research team led by Alex Nichol wrote in Point-E: A System for Generating 3D Point Clouds from Complex Prompts, published last week. "To produce a 3D object from a text prompt, we first sample an image using the text-to-image model, and then sample a 3D object conditioned on the sampled image. Both of these steps can be performed in a number of seconds, and do not require expensive optimization procedures."

If you were to input a text prompt, say, "A cat eating a burrito," Point-E will first generate a synthetic view 3D rendering of said burrito-eating cat. It will then run that generated image through a series of diffusion models to create the 3D, RGB point cloud of the initial image -- first producing a coarse 1,024-point cloud model, then a finer 4,096-point. "In practice, we assume that the image contains the relevant information from the text, and do not explicitly condition the point clouds on the text," the research team points out. These diffusion models were each trained on "millions" of 3d models, all converted into a standardized format. "While our method performs worse on this evaluation than state-of-the-art techniques," the team concedes, "it produces samples in a small fraction of the time." OpenAI has posted the projects open-source code on Github.

Artists Opposing AI Image Generators Use Mickey Mouse to Goad Copyright Lawsuits (dailydot.com) 149

Posted by EditorDavid on Sunday December 18, 2022 @02:48PM from the C-you-real-soon dept.

AI tools like DALL-E 2, Lensa AI, and Midjourney "can be told to create imagery in the style of a particular artist," notes this article in the Daily Dot.

Yet "The current legal consensus, much to the chagrin of many artists, concludes that AI-generated art is in the public domain and therefore not copyrighted." So... In response to concerns over the future of their craft, artists have begun using AI systems to generate images of characters including Disney's Mickey Mouse. Given Disney's history of fierce protection over its content, the artists are hoping the company takes action and thus proves that AI art isn't as original as it claims. Over the weekend, Eric Bourdages, the Lead Character Artist on the popular video game Dead by Daylight, urged his followers to create and sell merchandise using the Disney-inspired images he created using Midjourney.... "Legally there should be no recourse from Disney as according to the AI models TOS these images transcends copyright and the images are public domain."

Bourdages tweet quickly racked up more than 37,000 likes and close to 6,000 shares.

In numerous follow-up tweets, Bourdages generated images of other popular characters from movies, video games, and comic books, including Darth Vader, Spider-Man, Batman, Mario, and Pikachu.

"More shirts courtesy of AI," he added. "I'm sure, Nintendo, Marvel, and DC won't mind, the AI didn't steal anything to create these images, they are completely 100% original...."

Just two days after sharing the images, however, Bourdages stated on Twitter that he had suddenly lost his access to Midjourney.
The article notes that Bourdages reiterated his point in a later tweet. "People's craftsmanship, time, effort, and ideas are being taken without their consent and used to create a product that can blend it all together and mimic it to varying degrees."

DoNotPay Is Launching An AI Chatbot That Can Negotiate Your Bills (theverge.com) 21

Posted by BeauHD on Tuesday December 13, 2022 @06:30PM from the new-and-improved dept.

DoNotPay, the company that bills itself as "the world's first robot lawyer," is launching a new AI-powered chatbot that can help you negotiate bills and cancel subscriptions without having to deal with customer service. The Verge reports: In a demo of the tool posted by DoNotPay CEO Joshua Browder, the chatbot manages to get a discount on a Comcast internet bill through Xfinity's live chat. Once it connects with a customer service representative, the bot asks for a better rate using account details provided by the customer. The chatbot cites problems with Xfinity's services and threatens to take legal action, to which the representative responds by offering to take $10 off the customer's monthly internet bill.

This tool builds upon the many neat services DoNotPay already offers, which mainly allows customers can generate and submit templates to various entities, helping them to file complaints, cancel subscriptions, fight parking tickets, and much more. It even uses machine learning to highlight the most important parts of a terms of service agreement and helps customers shield their photos from facial recognition searches. But this is the first time DoNotPay's using an AI chatbot to interact with a representative in real time. The report notes that DoNotPay's bot is "built on top of OpenAI's GPT-3 API, the underlying toolset used by OpenAI's ChatGPT chatbot that tons of people have been playing around with to generate detailed (and sometimes nonsensical) responses."

What is ChatGPT, the AI Chatbot That's Taking The Internet By Storm 82

Posted by Slashdot Staff on Saturday December 03, 2022 @10:48PM from the its-alive dept.

A reader submits a report:
Artificial Intelligence (AI) research company OpenAI on Wednesday announced ChatGPT, a prototype dialogue-based AI chatbot capable of understanding natural language and responding in natural language. It has since taken the internet by storm, with people marvelling at how intelligent the AI-powered bot sounds. Some even called it a replacement for Google, since it's capable of giving solutions to complex problems directly," almost like a personal know-all teacher.

"We've trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests," OpenAI wrote on its announcement page for ChatGPT.

ChatGPT is based on GPT-3.5, a language model that uses deep learning to produce human-like text. However, while the older GPT-3 model only took text prompts and tried to continue on that with its own generated text, ChatGPT is more engaging. It's much better at generating detailed text and can even come up with poems. Another unique characteristic is memory. The bot can remember earlier comments in a conversation and recount them to the user. ChatGPT wrote a poem about Slashdot. And another one about Dogecoin.

Try ChatGPT for yourself here.

What Happened After Matt Taibbi Revealed Twitter's Deliberations on Hunter Biden Tweets? (wired.com) 377

Posted by EditorDavid on Saturday December 03, 2022 @03:46PM from the aftermath dept.

"Twitter CEO Elon Musk turned to journalist Matt Taibbi on Friday to reveal the decision-making behind the platform's suppression of a 2020 article from the New York Post regarding Hunter Biden's laptop," reports Newsweek.

"Taibbi later deleted a tweet showing [former Twitter CEO] Jack Dorsey's email address," adds the Verge, covering reactions to Taibbi's thread — and the controversial events that the tweets described: At the time, it was not clear if the materials were genuine, and Twitter decided to ban links to or images of the Post's story, citing its policy on the distribution of hacked materials. The move was controversial even then, primarily among Republicans but also with speech advocates worried about Twitter's decision to block a news outlet. While Musk might be hoping we see documents showing Twitter's (largely former) staffers nefariously deciding to act in a way that helped now-President Joe Biden, the communications mostly show a team debating how to finalize and communicate a difficult moderation decision.
Taibbi himself tweeted that "Although several sources recalled hearing about a 'general' warning from federal law enforcement that summer about possible foreign hacks, there's no evidence - that I've seen - of any government involvement in the laptop story."

More from the Verge: Meanwhile, Taibbi's handling of the emails — which seem to have been handed to him at Musk's direction, though he only refers to "sources at Twitter" — appears to have exposed personal email addresses for two high-profile leaders: Dorsey and Representative Ro Khanna. An email address that belongs to someone Taibbi identifies as Dorsey is included in one message, in which Dorsey forwards an article Taibbi wrote criticizing Twitter's handling of the Post story. Meanwhile, Khanna confirmed to The Verge that his personal Gmail address is included in another email, in which Khanna reaches out to criticize Twitter's decision to restrict the Post's story as well.

"As the congressman who represents Silicon Valley, I felt Twitter's actions were a violation of First Amendment principles so I raised those concerns," Khanna said in a statement to The Verge. "Our democracy can only thrive if we are open to a marketplace of ideas and engaging with people with whom we disagree."

The story also revealed the names of multiple Twitter employees who were in communications about the moderation decision. While it's not out of line for journalists to report on the involvement of public-facing individuals or major decision makers, that doesn't describe all of the people named in the leaked communications.... "I don't get why naming names is necessary. Seems dangerous," Twitter co-founder Biz Stone wrote Friday in apparent reference to the leaks.... The Verge reached out to Taibbi for comment but didn't immediately hear back.

Twitter, which had its communications team dismantled during layoffs last month, also did not respond to a request for comment.
Wired adds: What did the world learn about Twitter's handling of the incident from the so-called Twitter Files? Not much. After all, Twitter reversed its decision two days later, and then-CEO Jack Dorsey said the moderation decision was "wrong."
In other news, "Twitter will start showing view count for all tweets," Elon Musk announced Friday, "just as view count is shown for all videos." And he shared other insights into his plans for Twitter's future.

"Freedom of speech doesn't mean freedom of reach. Negativity should & will get less reach than positivity."

Slashdot Top Deals