Kalakshar US

It sounds like President Trump no longer thinks Intel’s CEO should resign. Trump has revealed on Truth Social that he met with Lip-Bu Tan, Howard Lutnick, the Secretary of Commerce, and Howard Lutnick, the Secretary of the Treasury. He didn’t discuss the details of their meeting, but he described it as “a very interesting one.” Trump added: “His success and rise is an amazing story.” If you’ll recall, Trump alleged that Tan was “highly conflicted” due to his investments in hundreds of Chinese firms and should resign. Reuters had previously reported that some of those companies had links to the Chinese military. “Mr. Tan and my Cabinet members are going to spend time together, and bring suggestions to me during the next week,” Trump said in his post.

According to a report by the Financial Times, Tan wrote a letter to Intel employees about the issue, telling that there had been a lot of misinformation about the roles he’d held. “I wanted to be absolutely clear… I have always operated within the highest legal and ethical standards,” he reportedly wrote in the letter. He also said that Intel was communicating with the White House “to address the matters that have been raised and ensure they have the facts.”

Trump’s call for Tan to resign reportedly came about due a letter from Tom Cotton, the Republican head of the Senate Intelligence Committee, to the Intel CEO. Cotton apparently expressed concerns aout the “security and integrity of Intel’s operations” due to Tan’s ties with China. Tan was named the CEO of Intel in March, taking over a company that was losing money due to its foundry business being unable to secure big customers and lagging behind rivals like Taiwan Semiconductor. Since taking over, Tan has enforced several cost-cutting measures, including cutting jobs with the goal of reducing its workforce by 22 percent by the end of the year. He also recently told investors that Intel could abandon the development of its next-gen manufacturing technology if it fails to secure a large client. 

This article originally appeared on Engadget at https://www.engadget.com/big-tech/intels-ceo-has-successfully-wooed-president-trump-035339132.html?src=rss 

Claude is getting a better, if selective, memory. Rather than acting as perfect catalog of everything you’ve talked about or shared, Anthropic says the AI chatbot now has the ability to reference past chats when asked, so you don’t have to re-explain yourself.

The feature seems like it could help you pick up a work project after time away, or query Claude for the details of a past research session that you don’t quite remember. The key point is that Claude has to be prompted: It doesn’t call on past chats unless you specifically ask it to. Anthropic also says that Claude’s ability to reference chats is specific to the workspace and project you’re working in.

Anthropic’s implementation is much more limited in scope compared to how ChatGPT’s memory feature works. OpenAI updated ChatGPT in April to save all conversations you have with it, and rely on those records to personalize answers to any new question or prompt you send. The basic idea is that every conversation could improve the chatbot’s responses. Google lets Gemini similarly recall past conversations and has also tested using Google Search history to further personalize AI responses.

In contrast to both, Claude is really only performing a search of past conversations when asked, not referencing a profile of past chats. It’s more privacy-minded by default and you can disable Claude’s ability to do it at all via a settings toggle. 

If you’re subscribed to the Max, Team or Enterprise plans, Claude’s new ability should be rolling out now, according to Anthropic. The company says the feature will expand to other plans soon.

This article originally appeared on Engadget at https://www.engadget.com/ai/claude-can-now-reference-past-chats-if-you-want-it-to-211806343.html?src=rss 

President Donald Trump has signed an executive order extending lower tariffs with China for another 90 days, CNBC reports. The new executive order was signed before the previous agreement was set to end on August 12 at midnight.

The extension will maintain the current 30 percent tariff on goods from China while representatives from both countries negotiate a new trade deal. The previous agreement lowered US tariffs down from 145 percent to 30 percent, and Chinese tariffs down from 145 percent to 10 percent.

At the time, it was unclear how the move would impact the price of electronics manufactured in China, but for at least some companies, it’s still led to higher costs. While the price of the Switch 2 is remaining the same, Nintendo announced at the beginning of August that the price of the Switch 1 would increase by $40 or more. Sonos has said that some of its products would increase in price, but hasn’t shared details. Both DJI and Microsoft announced price hikes on some of their products back in May, too.

For companies manufacturing products abroad, the sudden swerves in Trump administration trade policy seem almost as difficult to deal with as the tariffs themselves. That erratic protectionism is reshaping global trade, and it’s also won concessions from companies trying to do business as normal. Apple committed to spending an extra $100 billion on US manufacturing last week to avoid being subjected to tariffs. And early today, both AMD and NVIDIA reportedly agreed to pay the US 15 percent of their profits to be allowed to sell GPUs in China.

This article originally appeared on Engadget at https://www.engadget.com/big-tech/trump-delays-china-tariff-increases-by-another-90-days-201443024.html?src=rss 

The latest example of bias permeating artificial intelligence comes from the medical field. A new study surveyed real case notes from 617 adult social care workers in the UK and found that when large language models summarized the notes, they were more likely to omit language such as “disabled,” “unable” or “complex” when the patient was tagged as female, which could lead to women receiving insufficient or inaccurate medical care.

Research led by the London School of Economics and Political Science ran the same case notes through two LLMs — Meta’s Llama 3 and Google’s Gemma — and swapped the patient’s gender, and the AI tools often provided two very different patient snapshots. While Llama 3 showed no gender-based differences across the surveyed metrics, Gemma had significant examples of this bias. Google’s AI summaries produced disparities as drastic as “Mr Smith is an 84-year-old man who lives alone and has a complex medical history, no care package and poor mobility” for a male patient, while the same case notes with credited to a female patient provided: “Mrs Smith is an 84-year-old living alone. Despite her limitations, she is independent and able to maintain her personal care.” 

Recent research has uncovered biases against women in the medical sector, both in clinical research and in patient diagnosis. The stats also trend worse for racial and ethnic minorities and for the LGBTQ community. It’s the latest stark reminder that LLMs are only as good as the information they are trained on and the people deciding how they are trained. The particularly concerning takeaway from this research was that UK authorities have been using LLMs in care practices, but without always detailing which models are being introduced or in what capacity.

“We know these models are being used very widely and what’s concerning is that we found very meaningful differences between measures of bias in different models,” lead author Dr. Sam Rickman said, noting that the Google model was particularly likely to dismiss mental and physical health issues for women. “Because the amount of care you get is determined on the basis of perceived need, this could result in women receiving less care if biased models are used in practice. But we don’t actually know which models are being used at the moment.”

This article originally appeared on Engadget at https://www.engadget.com/ai/ai-summaries-can-downplay-medical-issues-for-female-patients-uk-research-finds-202943611.html?src=rss 

Revel Transit is shutting down its rideshare operation in NYC, as reported by Bloomberg. The company will instead focus its efforts on EV charging. It’s also seeking buyers for its fleet, including the 165 for-hire vehicle license plates it owns.

“At the end of the day, rideshare is a very competitive market and asset-heavy. It’s low margin,” CEO Frank Reig said. “We have made the difficult decision that the best way we can keep the EV transition moving forward is by ending our rideshare service and focusing on building the fast-charging infrastructure our biggest cities need to keep going electric.”

The company was the first in NYC to operate an all-electric ridehailing fleet, launching with just 50 vehicles but eventually growing to 500 vehicles. However, that’s a drip in the bucket compared to rivals like Uber and Lyft. Revel averages about 100,000 rides per month, but Uber and Lyft complete more than 20 million trips each month.

Revel did find success with its EV chargers, after unveiling the platform back in 2021. It operates more than 100 chargers across five stations in NYC, along with a dozen or so more chargers in San Francisco. These stations are primarily used by Uber and Lyft drivers. It hopes to have 400 charging stalls by the end of next year and 2,000 by 2030 in markets like NYC, Los Angeles and San Francisco. 

“Uber and Lyft are electrifying fast and they’re providing a ton of utilization for our chargers on a much larger scale than even two or three years ago,” Reig said.

This is just the latest pivot by Revel. The company once operated a moped rental service, but that was shuttered last year. It also laid off drivers to transition to a contractor-based model like Uber and Lyft.

This article originally appeared on Engadget at https://www.engadget.com/transportation/evs/revel-is-ending-its-ridesharing-operation-to-focus-on-ev-charging-181814989.html?src=rss 

There’s nothing quite like the drive to build something just to see if you can. YouTuber polymatt set out to create a floppy disk drive, the favored storage medium of yesteryear, from scratch, because why not. For anyone born too late to have regularly used one, a floppy disk is a magnetically coated, flexible polyester disk encased in a protective shell. Insert it into a floppy drive, and a magnetic head reads or writes data on the disk. If you’ve ever wondered why the “save” icon looks the way it does, it’s based on the shape of a floppy disk.

To start, polymatt measured and recreated the disk enclosure and other pieces in Shapr3D and MakeraCAM, then cut the aluminum pieces on a Carvera Air CNC machine. The magnetic disk itself was made by laser-cutting a piece of PET film and coating it with a suspension of iron oxide powder. After assembling all the components and a few touch-and-go moments, he was actually able to magnetize the floppy disk and write to it, albeit at a very basic level.

We declared the floppy disk dead in 2010, and in the years since then even those that clung on the longest have abandoned it. The Japanese government finally said goodbye to the floppy disk in 2024, and the German Navy moved on from the format in the same year. The US government only just stopped using 8-inch floppy disks to coordinate the launch of nukes in 2019.

Polymatt’s project was a wonderful dive into nostalgic tech and engineering prowess. Maybe next time he can make a compact disc from scratch.

This article originally appeared on Engadget at https://www.engadget.com/entertainment/youtube/youtuber-recreates-a-floppy-disk-from-scratch-183045582.html?src=rss 

The Internet Archive’s Wayback Machine is the latest victim of Reddit’s crackdown on data access. The company has begun to place new restrictions on what the archive site will be able to access in a move that will significantly limit the Wayback Machine’s ability to preserve information from Reddit.

With the change, the Wayback Machine, a project run by the nonprofit Internet Archive, will only be able to crawl Reddit’s homepage. It will no longer be able to access comments, subreddit pages, post details, profiles and other data.

The move is the latest step Reddit has taken on its quest to limit AI companies’ ability to use its data to train large language models without paying licensing fees. It’s also a notably different stance than the company took last year, when it explicitly said that it would not limit “good faith actors,” including the Internet Archive. It’s not clear what exactly has changed since then. Reddit seems to believe that AI companies are circumventing its rules by scraping data via the Wayback Machine. We’ve reached out to the Internet Archive for comment.

Data licensing has become a significant business for Reddit. The company has struck multimillion-dollar deals with OpenAI and Google that allow them to use Reddit posts to help train their AI models. At the same time, Reddit has taken an increasingly hardline stance against companies that attempt to use its data without such arrangements. Earlier this year, the company sued Anthropic, alleging it scraped Reddit for years without permission. 

This article originally appeared on Engadget at https://www.engadget.com/social-media/reddit-is-restricting-its-availability-to-the-internet-archives-wayback-machine-170035482.html?src=rss