Category: Technology

We’ve finally got new gadgets to review! This week, Cherlynn, Devindra and Engadget’s Sam Rutherford dive into the new 14-inch MacBook Pro with an M2 Max chip. Sure, it looks the same as before, but demanding users may appreciate the performance bump. Also, they discuss where the new M2 Pro-powered Mac Mini fits into Apple’s lineup. And of course, Cherlynn and Sam update us on everything they expect from Samsung’s Unpacked event next week. We’ll get the Galaxy S23, naturally, but rumors also point to new computers too.

Listen below or subscribe on your podcast app of choice. If you’ve got suggestions or topics you’d like covered on the show, be sure to email us or drop a note in the comments! And be sure to check out our other podcasts, the Morning After and Engadget News!

Subscribe!

iTunes

Spotify

Pocket Casts

Stitcher

Google Podcasts

Topics

MacBook Pro M2 Max review and Mac mini thoughts – 1:26

Samsung Unpacked 2023 preview – 13:02

Other news: Hacker leaks 2019 No Fly list – 27:14

Microsoft announces multibillion dollar investment in OpenAI days after layoffs – 33:45

Scientists found a colony of Emperor penguins after tracking poop markings on satellite images – 43:10

Formovie Theater UST projector and LG CineBeam projector reviews – 47:30

Ayaneo 2 handheld review: Like a Steam Deck, but fancier – 59:00

Pop culture picks – 1:06:14

Livestream

Credits
Hosts: Cherlynn Low and Devindra Hardawar
Guest: Sam Rutherford
Producer: Ben Ellman
Music: Dale North and Terrence O’Brien
Livestream producers: Julio Barrientos
Graphic artists: Luke Brooks

 

In a new study, University of Minnesota law professors used ChatGPT AI chatbot to answer graduate exams at four courses in their school. The AI passed all four, but with an average grade of C+. The University of Minnesota group noted ChatGPT was good at addressing “basic legal rules” and summaries, but it floundered when trying to pinpoint issues relevant in a case.

When faced with business management questions in a different study, the generator was “amazing” with simple operations management and process analysis questions, but it couldn’t handle advanced process questions. It even made mistakes with sixth-grade-level math – something other AI authors have struggled with.

If you’re unsure whether we’re ready for robot lawyers, you’re not the only one. DoNotPay is a free AI-powered chatbot that can draft letters and fill out forms for various legal matters, including appeals for parking tickets. Joshua Browder, the CEO of the New York startup, announced his company’s bot would represent a defendant fighting a traffic ticket in the courtroom on February 22nd. However, he received multiple jail threats from state bar prosecutors if he was to go through with his plan. DoNotPay is postponing its court case. Browder told NPR “The truth is most people can’t afford lawyers. This could’ve shifted the balance and allowed people to use tools like ChatGPT in the courtroom that maybe could’ve helped them win cases.”

– Mat Smith

The biggest stories you might have missed

BuzzFeed is the latest publisher to embrace AI-generated content

‘Dead Space’ is the new benchmark for video game remakes

Apple’s Studio Display drops to a record low price at Amazon

Amazon’s Blink Wired Floodlight Camera is now available for $100

Audi’s Activesphere EV concept is built for off-roading and augmented reality

OnePlus teases its first tablet ahead of next month’s official reveal

NYC wants all Uber and Lyft cars to be electric by 2030

Tesla Cybertruck won’t enter mass production until 2024

The electric pickup will only be available in small numbers this year.

During a conference call discussing Tesla’s latest earnings, company chief Elon Musk said mass production of the Cybertruck, its electric pickup, won’t begin until 2024. He still expects manufacturing to kick off “sometime this summer” but warned output would be “very slow” early on. Tesla unveiled the Cybertruck in 2019.

Continue reading.

Ayaneo 2 review: A Steam Deck Pro?

A nicer screen, way better performance, but double the price.

For people intrigued by the Steam Deck but want something with a bit more performance, the Ayaneo 2 is a great little (pricey) portable gaming PC, says Engadget’s Sam Rutherford. It features a newer Ryzen 7 6800U CPU that pumps out frame rates between 25 to 40 percent higher than the Steam Deck, while its sophisticated hall sensor joysticks deliver even more responsive controls. The Ayaneo 2 also features a seven-inch bezel-less LCD display with a higher 1,920 x 1,200 resolution and tons of connectivity, thanks to three USB-C ports. Crucially, it’s based on Windows 11 instead of SteamOS, which means it should run practically any game you can think of. The main downsides are somewhat short battery life (about two and a half hours on a charge), the lack of built-in touchpads and a starting price $450 higher than the Steam Deck. Yeah, that’s about double.

Continue reading.

DOJ says it disrupted a major global ransomware group

The agency claims the infiltration has thwarted over $130 million in ransom demands.

The US Department of Justice has spent months infiltrating and disrupting the Hive ransomware group, the agency announced on Thursday. The DOJ says Hive has targeted over 1,500 victims in 80-plus countries, extorting hundreds of millions of dollars in ransom payments. It first infiltrated Hive’s network in July 2022, providing over 300 decryption keys to Hive’s current victims and more than 1,000 keys to previous victims – preventing over $130 million in ransom payments.

Continue reading.

Hacker arrested for trying to sell personal information of nearly every Austrian citizen

The personal data includes full names, addresses and dates of birth.

Dutch authorities arrested a hacker for obtaining and trying to sell the personal information of nearly every Austrian citizen in May 2020 – and the defendant had also offered “similar data sets” from Italy, the Netherlands and Colombia. Authorities say the hacker posted in an online forum the nine million data sets, which police say consists of “registration data” residents must provide to authorities: full names, addresses and dates of birth – but no financial info. “Since this data was freely available on the internet, it must absolutely be assumed that these registration data are, in full or in part, irrevocably in the hands of criminals,” the police said.

Continue reading.

The best VR headsets for 2023

There’s never been a better time to jump into virtual reality.

Headsets have come a long way since the launch of the Oculus Rift and HTC Vive six years ago. The Meta Quest 2 has already been around for two years, and it’s proven to be a very capable portable VR experience. And if you’re looking for a more immersive experience, high-end PC headsets are getting cheaper (and there’s the new PS VR 2 to look forward to). Read on for our top picks.

Continue reading.

 

With the launch of the Galaxy Buds 2 last year, Samsung brought premium features like active noise cancellation (ANC) to its entry-level headphones, while keeping the same $150 price as the Galaxy Buds+. If that’s still too much money, you can now grab a pair at Amazon in multiple colors for just $95, for a savings of $55 or 37 percent. 

Buy Samsung Galaxy Buds 2 at Amazon – $95

The Galaxy Buds 2 received a solid 84 Engadget Review score thanks to a number of big improvements. The biggest is the addition of ANC, along with an adjustable ambient sounds mode. While the ANC doesn’t block external sounds as well as more expensive earbuds, it’s a big improvement over simple passive noise cancellation They also offer improved sound quality, with ample base and balanced sound. It also offers EQ presets and the ability to control features using the Galaxy Watch 4.

They’re 15 percent small and 20 percent lighter than the Galaxy Buds+, making them he company’s smallest and most comfortable earbuds to date. Battery life isn’t incredible at 5 hours (with ANC enabled), but the case holds an additional three full charges and it supports Qi wireless charging as well. Call quality is solid thanks to AI that helps reduce background noise.

All those features make the Galaxy Buds 2 a very solid buy, particularly at the $95 sale price that’s just off the all-time low. They’re designed to work best with Android devices, though so iPhone users will probably want to look elsewhere.

Follow @EngadgetDeals on Twitter and subscribe to the Engadget Deals newsletter for the latest tech deals and buying advice.

 

Apple has “halted the development” of its own WiFi chip that was meant to replace Broadcom’s in its devices “for a while,” according to Ming-Chi Kuo. The notable analyst explained in a Medium post that he’s basing this report on his latest survey of the semiconductor industry’s foundries, equipment, packaging and testing. If you’ll recall, Bloomberg reported earlier this month that the tech giant was working on its own wireless chips meant for devices slated for release in 2025. While Apple has yet to confirm the report, it’s not exactly hard to believe: The tech giant has been taking steps to design and manufacture more in-house components to lessen its reliance on outside companies. 

Kuo said Apple chose to devote most of its resources to developing its next-gen A-series and M-series processors instead. That way, it can ensure that the processors for its iPhones, iPads and MacBooks can enter production over the next couple of years. The analyst also explained that it’s riskier for Apple to use its own WiFi chips at a time when companies are switching their devices over to WiFi 6E. “Broadcom will be the biggest winner” in this situation, he said, since the iPhone 15 is expected to feature the new WiFi standard that allows access to the 6 GHz band.

While it’s unclear if Apple will ever release its own WiFi chip, Bloomberg’s Mark Gurman said the company’s wireless chipset ambition isn’t entirely dead. In Kuo’s Twitter thread about the report, Gurman chimed in and said that the tech giant is still working on a combined WiFi-Bluetooth chip. Gurman previously said that the tech giant is also working on a chip that combines Bluetooth, cellular and WiFi in a single component, but he didn’t say if that one is still under development.

When The Information published a piece last week that said Apple is working on a cheaper mixed-reality headset, one of its sources claimed that the device could use the company’s in-house Bluetooth and WiFi chipset. That would allow Apple to keep costs and the device’s final retail price low, since it wouldn’t have to deal with a third party company’s pricing demands. 

 

Google is rolling out a feature to Android that will add an extra layer of privacy when you browse websites in Incognito mode. The tech giant has revealed in a blog post (via The Verge) that the ability to lock Chrome Incognito sessions behind biometric authentication on mobile is currently making its way to Android users after debuting on iOS. When this feature is switched on, you will have to verify your identity through face or fingerprint authentication every time you return to an Incognito tab after exiting Chrome or navigating away to another app. 

Locked Incognito sessions first became available on Android in 2022 — over a year after Google started testing it for iPhones and iPads — but as an experimental feature you’d have to activate via a flag. Based on 9to5Google’s experience using it on Android last year, this feature will show you a gray screen with the Incognito logo at the center when you come back to a tab protected by the privacy feature. You’ll then have to tap “Unlock Incognito” to open biometric authentication or choose to verify your identity with your PIN code. 

While the feature isn’t available for us yet, you can check out if you already have access to it by tapping on the three-dot menu at the top right corner of Chrome. Open Settings and then go to Privacy and Security to toggle on “Lock Incognito tabs when you leave Chrome” if you already see it as an option. 

 

It might not be long before every ridesharing car in New York City is electric. Mayor Eric Adams has outlined an agenda that will require “high-volume for-hire” vehicles at Uber, Lyft and similar companies to be zero-emissions by 2030. There will be “no new costs” for drivers, the administration says. The initiative would build on the city’s plans to electrify its own fleet.

Adams didn’t detail how this transition would take place. The Vergenotes that the Taxi and Limousine Commission, which already regulates NYC ridesharing, would likely be responsible for implementing the EV strategy.

At least some companies are already onboard with the idea. Uber “applaud[s]” Adams’ plan, according to a statement, while Lyft says it’s “excited” to work with the city. It’s not a difficult target for them, however. Uber and Lyft were already planning to go completely electric by 2030. They also have programs in place to encourage EV adoption across the US, such as Uber’s rentals through Hertz as well as Lyft’s incentives. Pressure elsewhere might also leave services with little choice. California will require that most ride-hailing cars are EVs by 2030, for instance.

Drivers may face challenges, however. EVs are currently more expensive than their combustion engine counterparts, and workers may have trouble affording them even if the maintenance costs are ultimately lower. EV prices are declining, but it may be a while yet before they’re truly affordable to a driver base struggling to improve pay.

There’s also the question of infrastructure. A 2022 study led by the National Renewable Energy Laboratory estimated that NYC would need over 1,000 150kW fast charging stations to adequately power 20,000 rideshare and taxi cars, even if 15 percent of drivers could top up overnight. The mayor’s proposal would electrify “100,000-plus” rides — the city may need a major investment in charging facilities to make the switch.

 

There’s plenty of concern that OpenAI’s ChatGPT could help students cheat on tests, but just how well would the chatbot fare if you asked it to write a graduate-level exam? It would pass — if only just. In a newly published study, University of Minnesota law professors had ChatGPT produce answers for graduate exams at four courses in their school. The AI passed all four, but with an average grade of C+. In another recent paper, Wharton School of Business professor Christian Terwiesch found that ChatGPT passed a business management exam with a B to B- grade. You wouldn’t want to use the technology to impress academics, then.

The research teams found the AI to be inconsistent, to put it mildly. The University of Minnesota group noted that ChatGPT was good at addressing “basic legal rules” and summarizing doctrines, but floundered when trying to pinpoint issues relevant to a case. Terwiesch said the generator was “amazing” with simple operations management and process analysis questions, but couldn’t handle advanced process questions. It even made mistakes with 6th grade-level math.

There’s room for improvement. The Minnesota professors said they didn’t adapt text generation prompts to specific courses or questions, and believed students could get better results with customization. At Wharton, Terwiesch said the bot was adept at changing answers in response to human coaching. ChatGPT might not ace an exam or essay by itself, but a cheater could have the system generate rough answers and refine them.

Both camps warned that schools should limit the use of technology to prevent ChatGPT-based cheating. They also recommended altering the questions to either discourage AI use (such as focusing on analysis rather than reciting rules) or increase the challenge for those people leaning on AI. Students still need to learn “fundamental skills” rather than leaning on a bot for help, the University of Minnesota said.

The study groups still believed that ChatGPT could have a place in the classroom. Professors could teach pupils how to rely on AI in the workplace, or even use it to write and grade exams. The tech could ultimately save time that could be spent on the students, Terwiesch explains, such as more student meetings and new course material.

 

Dutch authorities arrested a hacker for obtaining and trying to sell the personal information of nearly every Austrian citizen in May 2020, according to Reuters. It includes almost nine million data sets, roughly lining up with Austria’s population.

The defendant, arrested in November in an Amsterdam apartment, was reportedly already known to international police. The 25-year-old defendant also offered “similar data sets” from Italy, the Netherlands and Colombia. Dutch police waited until now to announce the arrest to avoid hindering ongoing investigations.

Authorities say the hacker posted the information in an online forum. Police say the trove consists of “registration data,” essential info residents must provide to authorities. That includes their full name, address and date of birth — but not financial info, fortunately. Nevertheless, the police confirmed the material’s authenticity, adding that “since this data was freely available on the Internet, it must absolutely be assumed that these registration data are, in full or in part, irrevocably in the hands of criminals.”

 

New York City’s first flood-monitoring network is set to expand. Thanks to $7.2 million in funding from the city, the number of flood-prone areas FloodNet monitors with its sensors will increase from 31 to 500 across all five boroughs. The expansion is expected to start next month and take up to five years.

Data from the sensors is fed into a free web dashboard that residents, city agencies, researchers and anyone else can use to stay on top of and react to flood threats. The dashboard receives water-level readings from the sensors in real time. An interactive map shows where, when and how rapidly water is rising, whether that’s due to seawater surges at the coast or stormwater drains not being able to handle floods. The dashboard also includes historical data, which can help people to track the frequency and severity of floods over time.

Researchers from New York University, The City University of New York, Brooklyn College and the Science and Resilience Institute developed FloodNet. They had assistance from the mayor’s Office of Climate & Environmental Justice, the NYC Office of Technology & Innovation and neighborhood community groups.

FloodNet’s solar-powered sensors are low cost and open-source. They use ultrasound to measure changes in water levels and wirelessly transfer data to a gateway hub, which then sends the information to FloodNet’s servers and the dashboard.

Sea levels in the city have risen by a foot in the last century, according to the New York City Panel on Climate Change. They’re expected to increase by between another eight and 30 inches by around 2050, and between 15 and 75 inches by the end of the century. More detailed flood data can help city planners and others prepare for permanent water level rises, along with weather events like hurricanes that can quickly wreak havoc.

 

The US Department of Justice has spent months infiltrating and disrupting the Hive ransomware group, the agency announced on Thursday. The DOJ says Hive has targeted over 1,500 victims in more than 80 countries, extorting hundreds of millions of dollars in ransom payments.

Working with German and Netherlands law enforcement, the FBI seized Hive’s servers and websites, allegedly slowing the group’s ability to attack and extort new victims. It first infiltrated Hive’s network in July 2022, providing over 300 decryption keys to Hive’s current victims and more than 1,000 keys to previous victims — preventing over $130 million in ransom payments. The agency hasn’t announced any arrests. However, it’s still investigating the group, according toNBC News.

Hive used a ransomware-as-a-service (RaaS) model, where administrators (essentially the ringleaders) create ransomware strains with easy-to-use interfaces. The administrators then recruit affiliates who use the ransomware software to carry out the theft — and likely much of the risk.

For example, Hive would steal a victim’s data and encrypt their system. The affiliate would then demand a ransom in exchange for the decryption key and a promise not to publish the data. (Of course, it would frequently target the most sensitive data to apply maximum pressure.) If the victims pay, affiliates and administrators would split the ransom 80 / 20. Those unwilling to pay would find their data leaked on the web.

The US Cybersecurity and Infrastructure Security Agency (CISA) says Hive gained access through single-factor logins via Remote Desktop, VPNs, exploiting FortiToken (software-based access key) vulnerabilities and phishing emails with malicious attachments.

“Last night, the Justice Department dismantled an international ransomware network responsible for extorting and attempting to extort hundreds of millions of dollars from victims in the United States and around the world,” said US Attorney General Merrick Garland today. “We will continue to work both to prevent these attacks and to provide support to victims who have been targeted. And together with our international partners, we will continue to disrupt the criminal networks that deploy these attacks.” The FBI recommends victims contact their local FBI field office.