Dev.SN

"chromas" writes: from the teehee dept wiki.soylentnews.org: Foreword This is the documentation for the upcoming SoylentNews API that I'm working on. It is currently live but it's possible it may cease to be that way. In the event that it doesn't though, I want it documented so people can actually start building against it right off. Layout Code is arranged with methods (m=foo) being the highest order of operation, each able to perform several ops (m=foo&op=bar, m=foo&op=baz). Each op requires or allows different arguments, depending on what is being done. Methods The top level methods are as follows: Operations The operations for each method are as follows: user ops comment ops latest - returns the latest 50 comments without argument or all comments since the comment id supplied in the argument "since". if "since" is not numeric, the latest 50 are returned. if since is too high, an empty json array is returned post - posts a comment. sid, postersubj, posttype, reskey, and postercomment are required. sid may be numeric or in "14/08/07/1647258" form. preview=1 will return a json-encoded preview exactly as it would be included on the webpage for viewing. to reply to another comment, set pid = the comment to reply to's cid. I very much suggest you use a POST request here or it will be a huge pain to fill out postercomment properly and you will be limited by GET request max length. posttype is as follows: 1 Plain Old Text, 2 HTML, 3 Extrans (HTML tags to text), 4 Code reskey comes from a previous call to the m=comment&op=reskey op ten seconds are required between reskey creation and use so do a reskey creation call before allowing the user to enter data story ops latest - returns the latest 10 stories. optionally takes "limit" as an argument for the number of stories returned, which may range from 1-50. You may also specify "tid" as an argument. Behavior is however strange currently. If tid is the tid of a topic, you will get stories of that topic from every nexus; if tid belongs to a nexus, you will get stories of every topic within that nexus. This needs to be changed but it would require changing the main site as well. journal ops auth ops login - Logs you in, duh. Takes nick and pass as mandatory arguments. Currently returns the cookie value. DO NOT USE THE COOKIE VALUE. Use the proper cookie returned in the headers as this will change to a success/failure check only soon. You should also probably do this via POST for security reasons. Errata Related pages Example scripts Usage Syntax: ~api [%argument1%=%value1% %argument2%=%value2%] [/%element%] There can be as many argument=value pairs as you like, and the element is optional and corresponds to a key in the json structure ~api m=user op=get_uid nick=The Mighty Buzzard /uid The element may also contain forward-slashes that correspond to array levels, allowing you to drill down into an array structure to retrieve a specific piece of data. For example, to get the uid of the latest comment: ~api m=comment op=latest /0/uid Original Submission

janrinok writes: https://www.cnet.com/tech/services-and-software/florida-sues-openai-chatgpt-safety-lawsuit-news/ Florida Attorney General James Uthmeier at a podium with a sign reading "OpenAI accountability." Florida isn't buying OpenAI's promise to build safely, as the beginning of the complaint shows. The state's lawsuit against OpenAI and its CEO, Sam Altman, says they willfully ignored warnings, both from inside and outside of the company, about the many risks AI poses to its users. Florida alleges that OpenAI lied about ChatGPT's reliability, suitability for children and promotes prolonged use that leads to users' cognitive decline.  (Disclosure: Ziff Davis, CNET's parent company, in 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.) The lawsuit comes as Florida pursues a criminal investigation into whether ChatGPT played a role in last year's mass shooting at Florida State University that killed two people and injured six others. In that case, the shooter allegedly used ChatGPT to plan the attack, including advising on the type of weapon, the timing of the massacre and how to dispose of human bodies. At the time, OpenAI said: "Last year's mass shooting at Florida State University was a tragedy, but ChatGPT is not responsible for this terrible crime." OpenAI CEO Sam Altman is named in Florida's lawsuit. There are growing concerns about how ChatGPT and other chatbots can feed into people's violent actions and harmful delusions. Experts have found that chatbots like ChatGPT can struggle to push back on dangerous ideas and be so eager to please that they can provide factually incorrect information, a problem called sycophancy.  Another area of concern for legislators and tech watchdog groups is over OpenAI's data collection and privacy practices. Florida's complaint says that ChatGPT offers kids unfettered access to "harmful information" about eating disorders and self-harm. By concealing these risks and promoting ChatGPT as safe, OpenAI has misled Floridians and the general public with a dangerous product, the complaint says. OpenAI said in a statement that it believes minors need significant protections around AI and has worked to provide them to parents and teens. "Losing a child is the most devastating tragedy that can happen to a family and we know that no words can come close to addressing the pain of such a loss," an OpenAI spokesperson said. "We're committed to getting this right." While this is the first state-led lawsuit against OpenAI around child safety, numerous state governments are taking action around AI. California, Illinois and New York have created new laws to rein in how AI companies operate. Florida's lawsuit is a civil case, which would result in penalties (money) and court orders instead of criminal charges. Though it's unclear still how the financial penalty will play out, Meta and Google were recently ordered to pay $3 million after a jury found them guilty of creating addictive social media apps; in a separate case, Meta was ordered to pay $375 million on child exploitation charges. These cases deal with social media, not AI, but these legal strategies used against Big Tech could provide a legal roadmap going forward.  Despite a growing state and local backlash against AI, the Trump administration's newest AI plan shows it wants the federal government to be in charge of making the rules around the technology. The White House has been outspoken in its support for AI infrastructure projects, including the boom of data center construction projects across the US.  But experts warn that loosening regulations to let AI companies build faster could have disastrous effects on the environment, economy and society as a whole. Original Submission

janrinok writes: https://www.cnet.com/tech/services-and-software/anthropic-ipo-first-filing-ai-revenue/ An Anthropic representative didn't immediately respond to questions about the expected timing or valuation of the offering. The Claude-maker is one of three big tech firms expected to have initial public offerings this year amid what some call an "AI gold rush." SpaceX, the Elon Musk-owned rocket company that also includes the Starlink ISP, the AI lab xAI, and the social network now known as X, filed for an IPO in May. Anthropic's major rival, ChatGPT maker OpenAI, is expected to follow suit soon. The frenzied IPO race reflects the market's eagerness to cash in on its trillion-dollar bets, as AI companies rush to secure the massive funding needed to survive. The AI industry is capital-intensive, driven by the immense costs of maintaining the computing power required to train large language models, as well as the data centers, silicon and energy grids to keep them running.  Though Anthropic turned over voluntary documents for regulatory review, that doesn't guarantee a final decision on the IPO, and the company could still decide to not go public, according to Patrick Corrigan, a law professor at the University of Notre Dame. Based on typical SEC timelines, a public filing could be expected in a few weeks, with stock trading potentially starting in two to four months, Corrigan told CNET in an interview.  The AI industry has been a highly speculative landscape, where valuation is determined by a company's future potential rather than current profits. An online tracker of revenue and losses found that more than twice as much money has been spent on AI development as has been made back, pointing to billions of dollars in debt. The only major company to come out ahead is Nvidia, which makes the chips at the center of the AI gold rush.  Critics point out that AI companies have raised capital through manipulated accounting, using "annualized" revenue spikes and ignoring core costs to hide poor margins, thereby misleading investors.  "Their valuations are, at this point, so high that it's becoming increasingly impractical to raise more capital, and their investors are likely demanding some kind of liquidity event," said Ed Zitron, author of the Where's Your Ed At newsletter and host of the Better Offline podcast.  Anthropic said last week it raised $65 billion in a funding round that valued the company at $965 billion. The company has focused heavily on business customers and developers, with huge growth in 2026 spurred by its Claude Code programming tool, potentially putting it ahead of OpenAI in total value.  An IPO would be a big test of the business model of Anthropic and CEO Dario Amodei, seen here during the World Economic Forum in Davos, Switzerland, in January. But there are limits to what a company can raise on the private market, and those investors will want to make their money back with sizable returns. Going public and having stock bought not just by individual traders on Robinhood but also by institutional investors like insurance companies and pension funds could raise significantly more capital. These companies would also be able to use that stock to secure loans they might not be able to get now.  "The scale of it's eye-popping," said Rob Lalka, a business professor at Tulane University. "The fact that this company raised as much as they had raised, and they ran through quite a bit of money pretty quickly as they got here." The confidential submission means we don't yet have access to the documents required for a public stock sale, including a prospectus that details the business's operations and the risks and challenges it faces. That eventual public filing will have to include details on the company's financials and the numerous potential risks it faces. Transparency will also be key. "If they lie to investors, then the company could be held liable," Corrigan said.

I normally submit stories with no entry in the boxes:         "Your Name"         "Your Email or Homepage"Does this new dev version still do this? The preview looks positive, here goes, pressing "SubmitStory" now! Original Submission

JoeMerchant writes: Ho there, you're as welcome as can be. Original Submission