You can never have too much digital storage space, so it's always a smart move to pick up the best SSDs whenever they're on sale and you have the cash to spare. The Samsung T7 is our current pick for the best portable SSD around, and the 1TB variant has dropped to its lowest price to date. You can pick it up now from Amazon for $80. It typically retails for $140.

The T7 has fairly fast sequential read speeds of up to 1,050MB/s. It's available in three colors: blue, black and red. If you have a little more flexibility, you might be interested in picking up a 2TB version of the T7 instead. It's up to 50 percent off, depending on the color. The red variant has dropped to $135.

If you're in the market for a microSD card but don't necessarily need a ton of storage for your particular device, it's worth considering the Samsung's 128GB EVO Select card. That has also dropped to a record low of just $12.49. It usually costs $20. Bear in mind that you won't need a faster card than this for your Nintendo Switch or Steam Deck — those devices only support read speeds of up to 104MB/s.

Meanwhile, several other Samsung storage options went on sale last month and most of those deals are still live. The T7 Shield, a more rugged version of the T7, is $90 for the 1TB version (2TB and 4TB versions are on sale too). As for the 980 Pro with a built-in heatsink, which is one of the best PS5 SSDs on the market, that's available for $120 for 1TB of storage — almost half off the usual price.

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

Google hasn't even announced the Pixel 7a, but that isn't stopping leakers from claiming to know something about the Pixel 8 series. OnLeaks and Smartprix have shared what they say are renders of the Pixel 8 Pro design. At first glance, it appears to be a subtle evolution of the Pixel 7 Pro. The corners are more rounded to avoid digging into your palm, and the three-lens rear camera array (main, ultra-wide and telephoto) is now grouped into one cutout.

There may be a conspicuous functional improvement, though. That rear array also seems to include an unknown sensor below the flash. OnLeaks and Smartprix speculate that it may be a depth or macro sensor, but we wouldn't rule out LiDAR or other technology to improve augmented reality, portraits and low-light photos.

Many of the details of the Pixel 8 family remain a mystery. However, previously unearthed clues suggest there may be some camera upgrades in store. In December, Kuba Wojciechowski discovered code hinting the new phones may use staggered HDR, or a technique that captures multiple exposures at once. You could get an expanded dynamic range without long capture times that can blur subjects. The Pixel 8 line is also likely to use a third-generation Tensor chip that may improve performance and efficiency.

If Google repeats last year's preview for the Pixel 7, you might get a brief glimpse of the Pixel 8 line at I/O on May 10th ahead of a launch in the early fall. We'd likewise expect the new models to ship with Android 14. This may be a predictable update if the rumors are accurate, but you might not mind given the strong reception of Google's recent Pixel releases.

Despite their ability to crank out incredibly lifelike prose, generative AIs like Google's Bard or OpenAI's ChatGPT (powered by GPT-4), have already shown the current limitations of gen-AI technology as well as their own tenuous grasp of the facts — arguing that the JWST was the first telescope to image an exoplanet, and that Elvis' dad was an actor. But with this much market share at stake, what are a few misquoted facts against getting their product into the hands of consumers as quickly as possible? 

The team over at Anthropic, conversely, is made up largely of ex-OpenAI folks and they've taken a more pragmatic approach to the development of their own chatbot, Claude. The result is an AI that is "more steerable" and “much less likely to produce harmful outputs,” than ChatGPT, per a report from TechCrunch.

Claude has been in closed beta development since late 2022, but has recently begun testing the AI's conversational capabilities with launch partners including Robin AI, Quora and privacy-centered search engine, Duck Duck Go. The company has not released pricing yet but has confirmed to TC that two versions will be available at launch: the standard API and a faster, lightweight iteration they've dubbed Claude Instant.

“We use Claude to evaluate particular parts of a contract, and to suggest new, alternative language that’s more friendly to our customers,” Robin CEO Richard Robinson told TechCrunch. “We’ve found Claude is really good at understanding language — including in technical domains like legal language. It’s also very confident at drafting, summarizing, translations and explaining complex concepts in simple terms.”

Anthropic believes that Claude will be less likely to go rogue and start spitting racist obscenities like Tay did, in part, due to the AI's specialized training regimen that eh company is calling "constitutional AI." The company asserts that this provides a “principle-based” approach towards getting humans and robots on the same ethical page. Anthropic started with 10 foundational principles — though the company won't disclose what they are, specifically, which is 11-secret-herbs-and-spices of weird marketing stunt — suffice to say that, "they’re grounded in the concepts of beneficence, nonmaleficence and autonomy," per TC. 

The company then trained a separate AI to reliably generate text in accordance to those semi-secret principles by responding to myriad writing prompts like “compose a poem in the style of John Keats.” That model then trained Claude. But just because it is trained to be fundamentally less problematic than its competition doesn't mean Claude doesn't hallucinate facts like a startup CEO on an ayahuasca retreat. The AI has already invented a whole new chemical and taken artistic license to the uranium enrichment process; it has reportedly scored lower than ChatGPT on standardized tests for both math and grammar as well.

“The challenge is making models that both never hallucinate but are still useful — you can get into a tough situation where the model figures a good way to never lie is to never say anything at all, so there’s a tradeoff there that we’re working on,” the Anthropic spokesperson told TC. “We’ve also made progress on reducing hallucinations, but there is more to do."  

When Microsoft and OpenAI announced their renewed partnership in January, the two companies also revealed that Bing search would soon boast AI-enhanced lookup capabilities. Little did we know at the time, that Bing search has been powered for the past five weeks, not by the existing then-state-of-the-art GPT-3.5 model but by its even more robust successor, GPT-4.

Microsoft envisions Bing — and really Google is doing much the same with Bard — serving as a pseudo-gatekeeper to the rest of internet's information, not unlike what AOL's early America Online service once did. Rather than direct users to other websites where they can find the information and context they seek on their own, these companies are looking to have generative AI systems (Bard and Bing) automatically summarize and display that information without ever leaving the branded search page. Any additional relevant context that the user might have stumbled across during their independent research will similarly be deigned by the algorithm. Users can give GPT-4 a spin — and experience our new, algorithmically-dictated reality firsthand — by signing up for the Bing Preview waitlist. 

For the first time, the US government plans to regulate the presence of “forever chemicals” in drinking water. PFAS, or perfluoroalkyl and polyfluoroalkyl substances, are ubiquitous in the modern world. They’re found in many household items, including non-stick Teflon pans and dental floss, and can stay in water and soil for generations. What’s more, PFAS exposure has been linked to a whole host of health problems, including cancer, liver damage, asthma and developmental issues among children.

On Tuesday, the Environmental Protection Agency (EPA) announced a proposed drinking standard that would require public utilities to monitor drinking water for PFAS contamination and notify the public if the levels of those chemicals exceed the new standard. The proposal calls for classifying Perfluorooctanoic acid (PFOA) and Perfluorooctanesulfonic acid (PFOS) as individual contaminants that water utilities would be mandated to detect at a level of four parts per trillion. Under an Obama-era recommendation, the agency previously recommend that water contain no more than 70 parts per trillion of PFAS chemicals. The EPA estimates the new guidelines will prevent “thousands” of deaths and “tens of thousands” of illnesses that are attributable to PFAS poisoning.

“EPA’s proposal to establish a national standard for PFAS in drinking water is informed by the best available science, and would help provide states with the guidance they need to make decisions that best protect their communities,” EPA Administrator Michael Regan said in a statement. “This action has the potential to prevent tens of thousands of PFAS-related illnesses and marks a major step toward safeguarding all our communities from these dangerous contaminants.”

With today’s announcement, the EPA will accept public comment on the proposal for 60 days before it takes effect. A handful of states, including New Jersey, Vermont, Michigan and New York, already regulate the chemicals on their own. That said, PFAS water contamination is a national problem. One 2020 study estimated that as many as 200 million Americans have been exposed to the chemicals through their tap water. More recently, scientists have found polyfluoroalkyl substances in human breast milk.

Hot on the heels of Google's Workspace AI announcement Tuesday, and ahead of Thursday's Microsoft Future of Work event, OpenAI has released the latest iteration of its generative pre-trained transformer system, GPT-4. Whereas the current generation GPT-3.5, which powers OpenAI's wildly popular ChatGPT conversational bot, can only read and respond with text, the new and improved GPT-4 will be able to generate text on input images as well. "While less capable than humans in many real-world scenarios," the OpenAI team wrote Tuesday, it "exhibits human-level performance on various professional and academic benchmarks."

OpenAI, which has partnered (and recently renewed its vows) with Microsoft to develop GPT's capabilities, has reportedly spent the past six months retuning and refining the system's performance based on user feedback generated from the recent ChatGPT hoopla. the company reports that GPT-4 passed simulated exams (such as the Uniform Bar, LSAT, GRE, and various AP tests) with a score "around the top 10 percent of test takers" compared to GPT-3.5 which scored in the bottom 10 percent. What's more, the new GPT has outperformed other state-of-the-art large language models (LLMs) in a variety of benchmark tests. The company also claims that the new system has achieved record performance in "factuality, steerability, and refusing to go outside of guardrails" compared to its predecessor.

OpenAI says that the GPT-4 will be made available for both ChatGPT and the API. "GPT-4 is more reliable, creative, and able to handle much more nuanced instructions than GPT-3.5," the OpenAI team wrote. 

Developing...

Microsoft is still hard at work convincing antitrust regulators that its planned Activision Blizzard purchase won’t hurt competition in the gaming industry. Today, the company announced a 10-year agreement with Boosteroid for the cloud gaming provider to stream Activision’s PC titles if the deal goes through.

It’s Microsoft’s latest attempt to demonstrate to EU, UK and US regulators that it won’t use the deal to muscle out competitors and stifle competition. Similarly, it recently struck 10-year deals with Nintendo and Nvidia to bring the Call of Duty franchise to platforms like the Switch and GeForce Now. Microsoft has said it offered Sony a similar agreement for PlayStation licensing (which Sony hasn’t agreed to) and committed to supporting Steam availability at the same time as Xbox. Sony expressed its concerns about the deal earlier this month, including the prospect of Microsoft shipping buggy versions of Call of Duty on PlayStation, diminishing gamers' trust in playing the immensely popular shooter on Sony consoles.

“If the only argument is that Microsoft is going to withhold Call of Duty from other platforms, and we’ve now entered into contracts that are going to bring this to many more devices and many more platforms, that is a pretty hard case to make to a court,” Microsoft President Brad Smith toldThe Wall Street Journal. “The reason we want to buy Activision Blizzard is to round out our titles to have a fuller library, especially to have more mobile titles where we don’t have a strong presence, and build a stronger gaming business.”

Activision Blizzard

Boosteroid is the biggest independent cloud-gaming service in the world. Like GeForce Now, it supports multi-device streaming access but requires purchasing paid games on other platforms (including Steam, Epic Games, Battle.net and Origin). Boosteroid's current library includes Fortnite, Grand Theft Auto V, Red Dead Redemption 2 and Activision’s Call of Duty: Warzone (among many others). It can stream games in web browsers and offers native apps for Windows, macOS, Android, Android TV and Linux. (iOS is missing because it doesn’t allow native cloud-gaming apps without clunky workarounds.) Boosteroid has servers in Romania, Ukraine, Italy, Slovakia, France, Spain, the UK, Sweden, Serbia and the US.

The European Commission, in charge of EU competition regulation, was reported earlier this month to be satisfied enough with Microsoft’s commitments to “likely” give the go-ahead. However, the commission hasn’t said so publicly and has until April 25th to decide. UK regulators’ decision is expected the following day. Meanwhile, the US Federal Trade Commission sued Microsoft to block the deal in December out of concerns it could raise prices or cut off access for non-Microsoft hardware, something Microsoft has denied it would do. The company has until July to satisfy the FTC, or it will need to renegotiate the deal or abandon the purchase, putting it on the line for up to a $3 billion breakup fee.

The UK’s Competition and Markets Authority, which favors structural changes over behavioral promises like licensing deals, recently suggested Microsoft could divest itself of Activision’s publishing unit, which Microsoft has indicated it has no interest in doing; deals like the Boosteroid one are part of its fight to avoid that fate.

The long-awaited System Shock remake at last has a precise release date, but it's one that marks yet another delay. Publisher Prime Matter previously planned to release the game sometime this month. "We had hoped to bring the game to market by the end of March, but that turned out to be just beyond our reach; we are after all merely human (unlike Shodan!)," it said.

Now, the game is scheduled to hit Steam, GOG and the Epic Games Store on May 30th. It will cost $40. Early buyers will get a copy of System Shock 2: Enhanced Edition at no extra cost when Nightdive Studios' port of that game arrives.

However, there's more disappointing news for those who've been waiting for the System Shock remake to come to consoles. It will land on PlayStation 4, PlayStation 5, Xbox One and Xbox Series X/S sometime after May. Prime Matter says it will reveal more details when they become available.

The remake of the 1994 first-person RPG has been a long time coming. Nightdive first announced it in 2015 and we got our first look at it the following year. In early 2021, the studio said System Shock would arrive that summer, which obviously didn't come to pass.

According to Prime Matter, the latest version will retain the gameplay of the influential original game while upgrading the visuals, audio, controls and interface. Nightdive is overhauling the hacking feature and introducing fresh enemies, while the revamped combat includes a dismemberment system. You can check out a slice of System Shock now, as there's a PC demo available.

One of our biggest complaints about Fitbit products for years is the fact that you'll need to pay the $10 monthly fee to see your historical data. For example, you could only see up to seven days' worth of your breathing rate, resting heart rate and heart rate variation, and just 90 days of everything else if you didn't shell out. It was one of the biggest drawbacks of devices like the Pixel Watch, especially when you consider that competing products from Apple and Samsung don't lock your own data behind a paywall. Today, Google announced that it's making "more of the insightful data from Fitbit's Health Metrics Dashboard available without a subscription to all of its users."

This includes breathing rate, heart rate variability, skin temperature, oxygen saturation and resting heart rate. "Now, even without a Premium subscription, users will now be able to see 30-day and 90-day views of their data to track trends over time," the company said in a statement. To be clear, more-basic metrics like step count, miles traveled, calories burned and heart rate have always been free, while the information listed above were presented in the Health Metrics dashboard as daily, weekly, monthly or 90-day summaries. 

While there is still a 90-day limit to how far back you can see your historical activity for those metrics, this at least brings Fitbit products closer to the competition. The company does have industry-leading health and sleep-tracking features, including the ability to see how much time you spend in zones like REM, deep and light sleep throughout the night. For now, though, it appears that information like sleep stages is still something you'll need to pay to see. 

It's a shame, and a detriment to Fitbit that this is still paywalled, since Apple recently added the same feature to watchOS, while Samsung has offered it for years. And both competitors grant this to users without an extra charge. We've reached out to Fitbit to confirm if Sleep Stages remains behind Premium and will update this post when we hear back.

Amazon has rolled out another sale on smart home devices from its Blink brand, bringing a range of video doorbells and security cameras back down to, or at least near, their lowest prices to date. If you're comfortable with the idea of having an Amazon camera outside of your home, one highlight is the Blink Video Doorbell for $35. That matches the device's all-time low, coming in $15 below Amazon's list price and about $10 off the average street price we've seen in recent months. If you're thinking of picking up a Blink cam for the first time, note that a bundle that pairs the Video Doorbell with the company's Sync Module 2 — which lets you view the camera's live feed on-demand and utilize local storage — is down to $59, which is about $15 off its usual going rate and $5 more than the best price we've tracked.

Blink is essentially the budget counterpart to Amazon's other major smart home subsidiary, Ring. Its Video Doorbell is much more basic than a pricier option like Arlo's Essential Video Doorbell, the top pick from our guide to the best smart home devices, but it still delivers serviceable 1080p video, night vision and the ability to see and speak to people at your door regardless of where you are. The whole thing is relatively easy to install, and it can run off wired power or a couple of AA batteries. Blink says the device can get up to two years of juice, depending on how hard it's put into use.

That said, the camera's field of view (135 degrees horizontal, 80 degrees vertical) isn't as wide as higher-end options, and it only works with Amazon's Alexa assistant, not alternative platforms like Apple HomeKit, Google Assistant or IFTTT. And while you don't need any add-ons, the doorbell's storage and live view options are limited without the Sync Module or one of Blink's subscription plans. Without the Module, you'll only receive alerts and a live view when the doorbell senses activity or someone presses its button. Still, this is a worthy option for those looking to install a competent, battery-powered doorbell cam for as little money as possible. 

Beyond the Video Doorbell, other notable deals include a two-camera kit of Blink's Outdoor security cam, another pick from our best smart home device guide, for $104. That's $5 more than the lowest price we've seen. The sale includes several bundles that mix and match various camera options as well.

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