President Donald Trump has pardoned Binance founder Changpeng Zhao, the White House said. Zhao pleaded guilty to federal money laundering charges in 2023 and he was sentenced last year to four months in prison. He was released in September 2024. 

As part of his plea deal, Zhao stepped down as CEO of Binance and he was banned from having any involvement with the company for three years. Both Zhao and Binance reportedly submitted formal applications for pardons by August this year.

Trump “exercised his constitutional authority by issuing a pardon for Mr. Zhao, who was prosecuted by the Biden Administration in their war on cryptocurrency,” White House press secretary Karoline Leavitt said in a statement. “The Biden Administration’s war on crypto is over.”

The Wall Street Journal notes that the pardon could pave the way for Binance to start doing business in the US again. The company was barred from operating there after pleading guilty to violating money laundering laws in 2023.  Binance officials are said to have met with Treasury Department representatives this year in an attempt to reduce US oversight of the company.

Binance is involved with the Trump family’s World Liberty Financial cryptocurrency business — a venture that has padded the president’s pockets. For one thing, it bolstered the growth of USD1, a World Liberty cryptocurrency that’s pegged to the dollar. Binance received a $2 billion investment this spring and that was paid in USD1. According to CNBC, World Liberty has generated around $4.5 billion since last year’s presidential election.

FTX founder Sam Bankman-Fried has also reportedly been angling for a Trump pardon. Bankman-Fried was sentenced to 25 years in prison in 2024 after being found guilty of fraud and conspiracy to commit money laundering.

Crypto billionaire Justin Sun said last November that he’d invested $30 million into World Liberty (a figure that later rose to $75 million). In February, the Securities and Exchange Commission dropped a case against Sun. The agency had charged him in 2023 with alleged violations of securities laws. 

This article originally appeared on Engadget at https://www.engadget.com/big-tech/binance-founder-changpeng-zhao-lands-a-trump-pardon-174929498.html?src=rss

What if Clippy were powered by AI? That seems to be the pitch behind Microsoft’s new “expressive, customizable and warm” face of Copilot’s voice mode. The friendly blob listens, reacts and changes color in response to user interactions.

Microsoft sees Mico as an answer to what an “AI companion” looks like. The “optional visual presence” aims to listen and support without kissing ass. “It will push back on you sometimes, but always respectfully,” Microsoft AI CEO Mustafa Suleyman  wrote in a blog post.

But don’t take my word for it. Get ready for the most exciting 39 seconds of your day, as you watch Mico silently spin and shift hues.

Clippy — I mean, Mico — is also part of a new Copilot feature called Learn Live. The student-focused voice mode will have Mico act as a Socratic tutor that “guides you through concepts instead of just giving answers.” Its tools will include questions, visual cues and interactive whiteboards.

The Verge reports that Mico is only available in the US, UK and Canada at launch. The character is now being enabled by default for Copilot’s voice mode. But you can turn it off if talking to fictional characters isn’t your thing.

This article originally appeared on Engadget at https://www.engadget.com/ai/mico-is-microsofts-clippy-for-the-ai-age-174524597.html?src=rss

Assasin’s Creed Shadows is coming to the Nintendo Switch 2 on December 2, Ubisoft announced today in its fall roadmap update. The Switch 2 drop will include all of the game’s updates aside from the recent Claws of Awaji expansion, which will be added in 2026. It’ll also support cross-progression through Ubisoft Connect, allowing players to continue a current game on the new platform. The Switch 2 version of Assassin’s Creed Shadows will support a unique feature, too: touchscreen interactions in the menus, including the world map, store and hideout.

Ubisoft’s fall roadmap for Assassin’s Creed Shadows has three major dates. On October 28, title update 10 will target the parkour system, adding an Advanced Parkour option and reintroducing the directional catch ledge mechanic, among other tweaks. This update will also add a new Animus themed activity called Corrupted Castles. Then, on November 25, Ubisoft will roll out a third story drop and a surprise collaboration. The new quest is called “A Puzzlement” and it’s sillier in tone than Shadows itself, “much more akin to something like AC Odyssey,” according to developers. In it, Naoe will learn a version of Yasuke’s war kick ability, and Yasuke will discover how to sneak-stun enemies. On December 2, the Switch 2 release caps off the roadmap.

The Claws of Awaji expansion brought more than 10 hours of new content to Assassin’s Creed Shadows in September, introducing new weapons, abilities, gear, enemy factions, bosses and story beats, plus a fresh region to explore. This will all come to the Switch 2 version of the game later in 2026.

Assassin’s Creed Shadows debuted on PC, Mac, PlayStation 5 and Xbox Series X/S on March 20, 2025, and it’s a welcome bright spot in Ubisoft’s library. The studio recently released Star Wars Outlaws on Switch 2 and the consensus seems to be that it runs surprisingly well.

This article originally appeared on Engadget at https://www.engadget.com/gaming/nintendo/assassins-creed-shadows-will-hit-switch-2-on-december-2-172338735.html?src=rss

UN Secretary-General Antonio Guterres says it is now clear that efforts to cap global warming at 1.5C above pre-industrial levels will fail in the short term. AFP: Before next month’s COP30 climate summit in Brazil, Guterres said going beyond 1.5C would result in “devastating” yet predictable impacts. “One thing is already clear: we will not be able to contain the global warming below 1.5C in the next few years,” Guterres said at the UN’s World Meteorological Organisation (WMO) weather and climate agency in Geneva.

“Overshooting is now inevitable. Which means that we’re going to have a period, bigger or smaller, with higher or lower intensity, above 1.5C in the years to come.” However, if leaders start taking the problem seriously by driving towards net zero greenhouse gas emissions, “the 1.5 still remains — according to all the scientists I met — possible before the end of the century.”

Read more of this story at Slashdot.

An environmental watchdog group has suggested that millions of tons of discarded electronics from the US turn up in Asia and the Middle East each month, according to a report by ABC News. This has created a “hidden tsunami” of e-waste.

The Seattle-based Basel Action Network (BAN) conducted a two-year investigation into the matter and reportedly discovered at least ten US companies that exported used electronics to countries like Vietnam and Malaysia. This waste includes stuff like discarded phones and computers, which is an issue due to integrated toxic metals like lead, cadmium and mercury.

All told, around 2,000 containers of e-waste leave the US each month, totaling around 33,000 metric tons. The companies behind these shipments are described as “e-waste brokers,” as they don’t recycle the waste themselves. Rather, they ship it off for someone else to deal with. These entities have names like Corporate eWaste Solutions, Semsotai, First America Metal Corp. and PPM Recycling.

The ten companies named in the report allegedly exported e-waste valued at over $1 billion between January of 2023 and February of 2025. Trade could top out at more than $200 million each month across the industry.

Semsotai has responded to the allegations, saying that it doesn’t export scrap and only specializes in working components intended for reuse. It also accused BAN of bias. PPM Recycling accused BAN of exaggerating shipment volumes. Most of the companies named in the report, however, have chosen not to comment at this time.

The report also indicates that US e-waste shipments allegedly made up six percent of all US exports to Malaysia between the years 2023 to 2025. “Malaysia suddenly became this mecca of junk,” said Jim Puckett of BAN. Containers have also been tracked to Vietnam, Indonesia, Thailand, the Philippines and the UAE.

It’s worth noting that global e-waste is growing five times quicker than recycling programs. The world produced 62 million metric tons of electronic waste in 2022, a number that’s expected to climb to 82 million by 2030. This is according to the United Nations’ International Telecommunication Union and its research arm, UNITAR

Most countries around the world have banned these kinds of imports under the Basel Convention, which is an international treaty addressing hazardous waste trade. The US is the only industrialized nation that hasn’t ratified the treaty. This means that America, which is the second largest producer of e-waste in the world, is not bound by any of the rules set forth in the Basel Convention.

This article originally appeared on Engadget at https://www.engadget.com/big-tech/american-e-waste-is-causing-a-hidden-tsunami-of-junk-in-southeast-asia-171344351.html?src=rss

Australia’s First Peoples may or may not have hunted the continent’s megafauna to extinction, but they definitely collected fossils.

A team of archaeologists examined the fossilized leg bone of an extinct kangaroo and realized that instead of evidence of butchery, cut marks on the bone reveal an ancient attempt at fossil collecting. That leaves Australia with little evidence of First Peoples hunting or butchering the continent’s extinct megafauna—and reopens the question of whether humans were responsible for the die-off of that continent’s giant Ice Age marsupials.

Fossil hunting in the Ice Age

In the unsolved case of whether humans hunted Australia’s Ice Age megafauna to extinction, the key piece of evidence so far is a tibia (one of the bones of the lower leg) from an extinct short-faced kangaroo. Instead of hopping like their modern relatives, these extinct kangaroos walked on their hind legs, probably placing all their weight on the tips of single hoofed toes. This particular kangaroo wasn’t quite fully grown when it died, which happened sometime between 44,500 and 55,200 years ago, based on uranium-series dating of the thin layer of rock covering most of the fossils in Mammoth Cave (in what’s now Western Australia).

Read full article

Comments

Sora, OpenAI’s short-form AI video platform, has been available for iPhones for the past few weeks. Users have spent that time generating hyper-realistic vertical videos that, despite their watermarks, might already be fooling people on the internet.

I, for one, see nothing good about a tool that can make these types of videos from simple prompts, perhaps outside the sheer novelty alone. But the app is undeniably popular, and, as expected, OpenAI is continuing to “improve” it. On Wednesday, the head of Sora, Bill Peebles, shared a roadmap of updates coming to the app, from cameos for your pets to video editing tools.

Sora will soon let you make cameos of anything you want

Peebles says that “character cameos” are on their way in the coming days. When this feature drops, you’ll be able to make cameos of just about anything you want, including your pets (Peebles suggests your dog or guinea pig), stuffed toys, as well as AI-generated characters Sora may have produced in your past videos.

Cameos are a defining, and controversial, Sora feature. It lets you scan your likeness into the app so you can create videos featuring yourself. Not only that, you can remix videos with cameos of anyone else who has scanned themself into Sora—assuming their permissions allow you to use their cameos. It’s impressive tech, but it raises major ethical concerns, considering you can make someone else do or say things they never signed off on. Now, apparently, you’ll also be able to do the same with the family cat or your childhood stuffed animal.

Following these additional cameo options, Peebles says you’ll be able to see “the latest trending cameos in real time.” I’m guessing the idea is to browse the most popular cameos on the platform, and, if you want to, add them to your own videos.

Sora is also getting basic video editing tools. According to Peebles, this starts with the ability to stitch multiple clips together, though other “powerful new features” will follow. Peebles says the Sora team is also working on reducing “excessive moderation,” which he acknowledges is “super annoying” to users, as well as improving the app’s performance.

Speaking of apps, Android users will soon have their chance to try Sora. Right now, the app is iOS-only, but Sora’s Android version is “actually coming soon,” Peebles says.

A reminder to be wary of what you see on the internet

As critical and cynical as I am about Sora and services like it, I can see the utility here. I imagine people will get a kick out of being able to put their pets in any situation they want. And if they want to make fried eggs sing a song, or a stuffed animal say “hello,” more power to them.

But my main concerns with this app are not assuaged by any of these updates: Sora still lets you generate convincing videos with a simple prompt, with the only safeguard a watermark that’s easy to remove if you know where to look. People were already falling for fake videos on the internet before these tools came along; now, it’s becoming way too difficult to know whether what you’re watching is actually real or not. If you ask me, you’re better off assuming all of it is fake.

Disclosure: Lifehacker’s parent company, Ziff Davis, filed a lawsuit against OpenAI in April, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.

Back in August, Anthropic made Claude capable of remembering past conversations. With the update, people could reference specific chats, so that they wouldn’t need to repeat themselves when revisiting a topic. Today, the company has begun out a new, enhanced memory feature set, with the included improvements coming to all paying users.  

Plenty of chatbots, including ChatGPT and Gemini, can remember past conversations, but Anthropic believes its implementation has a few legs up on the competition. For one, Claude will learn your preferences and work patterns over time, which Anthropic says will translate to the chatbot getting better at understanding how you work. Additionally, the company claims Claude is “fully transparent” about its memory, meaning users will see an “actual synthesis” of what it has recorded over time, instead of “vague summaries.” If you want to edit its memory, you can do so through conversation. 

At the same, Anthropic has made it easy to compartmentalize the data Claude collects. When using the Projects feature to group conversations together, the chatbot will create a distinct memory space for each grouping. In this way, information Claude has saved from your work conversations won’t bleed over to your personal chats, for example. If you’re coming from ChatGPT or Gemini, Anthropic has made it possible to import saved memories from those chatbots to Claude. You can also export any tidbits of context Claude saves to other AI platforms. 

Ahead of today’s announcement, Anthropic notes it conducted extensive testing to determine if Claude’s new capabilities would lead to greater sycophancy and more harmful conversations. “Though this testing, we identified areas where Claude’s responses needed refinement and made targeted adjustments to how memory functions,” the company said. “These iterations helped us build and improve the memory feature in a way that allows Claude to provide helpful and safe responses to users.” 

Max subscribers can enable Claude new memory capabilities starting today, with availability for Pro users to follow in the coming days. The feature is fully optional, and won’t be turned on unless you toggle it through the settings menu.

This article originally appeared on Engadget at https://www.engadget.com/claude-can-now-compartmentalize-as-part-of-a-major-memory-upgrade-170000194.html?src=rss

People are evidently never buying quite enough stuff from Amazon to keep the company entirely happy, and it’s calling on AI once again to push indecisive shoppers into locking down the purchase they’ve been eyeing up. The new tool, which Amazon calls Help Me Decide, gives shoppers in the US personalized recommendations of products they should buy by analyzing their browsing history, searches and preferences.

It’s designed to “help” customers who have been looking at a number of products in a particular category, such as wireless headphones, to decide which one best suits their needs. The Help Me Decide button will pop up on a product detail page when it detects that you’ve been browsing for a while without making a final choice. If you choose to tap for AI assistance, it will pull together all the information it can find on your relevant shopping history and recommend the product it deems the right choice for you. It also recommends an alternative upgrade pick and a similar product for those on a budget.

Help Me Decide can also group together related searches. Amazon uses the example of the tool recommending an all-season tent for four people based on you previously looking for adult and kids’ sleeping bags that keep you warm, camping accessories and children’s hiking boots. The recommendation it chooses includes an explanation of why it’s the best pick for you based on its features and your previous purchases, and pulls in customer reviews to back it up. This suggests that how useful the recommendations are will ultimately come down to how much you take notice of customer reviews.

When searching for products to recommend to you, Help Me Decide leverages Amazon’s Bedrock and SageMaker machine learning platforms, as well as its OpenSearch tool, to marry up all the different factors it takes into consideration. It follows the introduction of the Interests tool earlier this year, which uses AI to generate shopping results based on your natural language prompts. Back In May, the company also started experimenting with AI-generated hosts that can summarise products for you before you buy them, again relying heavily on customer reviews for its information.

Help Me Decide is live in the US now and can be found in the Amazon app (iOS and Android) and mobile browser. If you tap “Keep shopping for” it should show up, and will do the same on a product detail page after you’ve looked at a number of products in a related category.

This article originally appeared on Engadget at https://www.engadget.com/ai/amazon-calls-on-ai-once-again-with-its-new-help-me-decide-shopping-tool-164516673.html?src=rss

From the outside, Counter-Strike 2 looks a lot like a game that’s primarily about shooting people. For millions of players, though, the game is more about collecting and/or buying rare in-game loot and flipping it for what can be very significant sums on the Steam Marketplace.

Wednesday night, Valve sent that multi-billion-dollar market into turmoil as part of a so-called “small update.” Now, players can use the game’s “Trade Up contracts” to exchange five common, “Covert” items (also known as “reds”) for the kinds of knives and gloves that have until now been much harder to obtain.

That “small update” has unsurprisingly had an immediate and sharp impact on the Marketplace price for those items. One rare knife that sold for over $14,000 less than 24 hours ago has seen its minimum price plummet over 50 percent as of this writing, according to the trackers at Pricempire. Meanwhile, the median sale price for a common P90 Asimov gun on the Steam Marketplace shot up from $10 on Wednesday to well over $100 as of this writing.

Read full article

Comments

An anonymous reader shares a report: With gingko “memory-boost tinctures,” fennel “tummy-soothing syrups” and “citrus-immune gummies,” AI “slop” has come for herbalism, a study published by a leading AI-detection company has found. Originality.ai, which offers its tools to universities and businesses, says it scanned 558 titles published in Amazon’s herbal remedies subcategory between January and September this year, and found 82% of the books “were likely written” by AI.

“This is a damning revelation of the sheer scope of unlabelled, unverified, unchecked, likely AI content that has completely invaded [Amazon’s] platform,” wrote Michael Fraiman, author of the study. “There’s a huge amount of herbal research out there right now that’s absolutely rubbish,” said Sue Sprung, a medical herbalist in Liverpool. “AI won’t know how to sift through all the dross, all the rubbish, that’s of absolutely no consequence. It would lead people astray.”

Read more of this story at Slashdot.