ArduinoMania

Repairability website iFixit has published its teardown of the iPhone 15 Pro Max, and the results are a mixed bag. Local repair shops still have to deal with the company's software-restricted “parts pairing” requirement, which means they need to order official components directly from Apple and get on the phone with a company employee before iOS will accept individual part replacements.

On the positive side, iFixit praised Apple for returning to a “dual-entry” removable glass back cover with the iPhone 15 Pro models — a feature that debuted with the standard iPhone 14 line last year. “This is a win for consumers as back glass repairs have been outrageously expensive on the high-end models until now, costing as much as $550,” iFixit said in its teardown video.

iFixit also examined the phone’s titanium frame and came away less than impressed. While noting that titanium is dirtier to produce than stainless steel and aluminum (mocking Apple’s “Mother Nature” skit in its launch event), the site also said the material scratches easily. “Unfortunately for the cool factor, we found that the color on the titanium shell scratches easily, a process that is only satisfying under the magnificent magnification of the microscope,” the teardown video said. “I could scratch this thing up all day.”

Elsewhere, iFixit found that the iPhone 15 Pro Max’s logic board appears to be the same as the one in the iPhone 15 Pro, and you have to remove the speaker and Taptic Engine to access the battery-removal tabs. Interestingly, the website also noted that the main and wide camera sensors on the iPhone 15 Pro Max appear identical to those on the iPhone 14 Pro Max, suggesting the “Tetraprism” periscope lens, which enables 5x optical zoom, is the only hardware-based camera update this year. “Any improvement in image quality has more to do with a new A17 SoC than the camera hardware itself,” iFixit said.

Dinging Apple for parts pairing appears primed to stand as a primary focus of iFixit’s Apple teardowns from now on. The repair advocacy website views it as significant enough of a problem to have lowered the iPhone 14’s repairability score from 7 out of 10 to 4 out of 10 nearly a year after launch because of it. “And as we’ve now come to expect, each year brings new parts pairing issues and bugs,” the video said. “This year’s edition is the LiDAR sensor, which now crashes if the sensor is swapped out. Calibration issue or not, these bugs need to be fixed, or else they might as well be paired with the logic board with a tiny Apple warning saying, ‘Hey, this phone is property of Apple.’”

Due to the parts pairing requirement, iFixit gave the iPhone 15 Pro Max a mere 4 out of 10 repairability score. “This phone won't accept salvaged parts, it complicates at-home repair, and it won't be any fun for your local repair tech,” the website said.

Bayonetta director Hideki Kamiya is leaving PlatinumGames, after helping to found the company back in 2006 when it was called Seeds Inc. Kamiya was recently promoted to VP, so this move comes as a slight surprise. He recently said on social media that it was "by no means an easy decision to make."

Kamiya still has a couple of weeks on his post, as he officially exits the company on October 12. As for the why of it all, he wrote that the move “came after a lot of consideration based on my own beliefs," but didn't offer further explanation. He says he’s still going to make games in his “Hideki Kamiya way”, but hasn’t announced if he’s heading to another company, starting his own or even just planning to tinker away in a garage somewhere. He’s only 52, so a complete retirement is highly unlikely.

Kamiya’s mark on gaming is massive. Most recently, he was the supervising director of the critically-acclaimed Bayonetta 3. During his more than 15 years at PlatinumGames, Kamiya worked on classics like the original Bayonetta, the Wii U and Switch cult hit The Wonderful 101 and the action-heavy Astral Chain, among others. He was staffed at Capcom and its spin-off studio Clover before founding PlatinumGames, helming Resident Evil 2, Viewtiful Joe and Ōkami.

Kamiya has been hard at work these past few years on a superhero title internally referred to as Project GG. He was lead director on the “heroic” game and the company has marketed it as a conclusion to his superhero trilogy, joining Viewtiful Joe and the Wonderful 101. PlatinumGames hasn’t announced if the game’s still coming or if it will poof into vaporware with Kamiya’s departure.

Resident Evil Village is a haunting horror romp starring a very tall and elegant vampire lady (and some other monsters, sure), and it's heading to iPhone 15 Pro and Pro Max on October 30. It'll hit the M1 and M2 models of the iPad Pro and iPad Air on the same day. The base game will cost $40 and its Winters’ Expansion DLC will be an additional $20.

Resident Evil Village originally came to PC, PlayStation 4, PS5, Xbox One and Xbox Series X/S in 2021, and it became a cultural touchstone for its monster-mashing storyline. The game includes werewolf creatures, a mutant fish man, a murderous cult leader and festering, zombie-like enemies, though its breakout star was Countess Alcina Dimitrescu. She's an exceptionally tall, undead, razor-fingered villain who leads a trio of vampiric daughters, and she's simply fantastic.

Village landed on Mac in 2022. Apple revealed the iPhone and iPad versions during its annual iPhone event on September 12, 2023, but it didn't share a release date at the time. Capcom provided the date on its site this week. The Resident Evil 4 remake, which landed on PC and consoles this year, is also due to hit Apple's mobile devices in 2023, but no date has been confirmed just yet.

Other games coming to the iPhone 15 Pro — thanks to the new A17 Pro chipset — include Death Stranding and Assassin's Creed Mirage. Death Stranding is due out this year, while Mirage is scheduled to hit in early 2024.

It's always nice to have some options when it comes to playing games on a fresh console, so for the next month Sony is giving away a free title to anyone who purchases and activates a new PS5.

Dubbed the Upgrader Program, Sony's latest initiative to entice potential PS5 buyers is refreshingly straightforward. In order to get a free game, users will need to purchase and activate their console before 11:59PM PT on October 20th. Once that's done, you can just go to the PlayStation Store and redeem a specific title by tapping on a banner for this offer. That said, if you don't already have an existing PSN account, you will need to make one as the free games come in the form of a digital download. 

The other nice thing is the selection of free titles includes a number of high-profile releases from the past few years. Here's the full list of currently redeemable games:

• Marvel’s Spider-Man: Miles Morales

• Marvel’s Spider-Man: Remastered

• God of War Ragnarök

• Horizon Forbidden West

• Ghost of Tsushima Director’s Cut

• Ratchet & Clank: Rift Apart

•  Demon’s Souls

• The Last of Us Part I

• Sackboy: A Big Adventure

• Returnal

• Uncharted: Legacy of Thieves Collection

• Death Stranding: Director’s Cut

Unfortunately, the Upgrader Program is only valid for owners in the US and it seems that if you have recently purchased a PS5 and activated it prior to September 23, you may not be eligible for the new offer. But for those who are able to take part, this is a great way to kick off your PS5 game collection with basically no strings attached.  

It's only been a few weeks since Analogue released a glow-in-the-dark Pocket console, but the manufacturer is already gearing up for yet another limited edition launch that could evoke memories of your youth. Analogue will start selling transparent Pockets, which are reminiscent of clear Game Boy Color consoles, on September 29 at 8 AM PT/11 AM ET.

You will have seven transparent colors to choose from: clear, smoke, red, blue, orange, green and purple. The retro gaming console will set you back $250 each — that means they cost $30 more than the basic versions, which are out of stock at the moment. These consoles are only available in limited quantities, though, and Analogue told us that they will never be sold again. If you're interested, you may want to online at the exact time the model will be available to check out, because the glow-in-the-dark edition sold out in mere minutes. 

Analogue's Pocket handheld can play Game Boy and Game Boy Advance cartridges out of the box and even support original accessories for the Nintendo console. You can also connect it directly to a Game Boy if you still have one for multiplayer gaming. But Game Boy cartridges aren't the only ones you can play again with a Pocket: Analogue also sells Game Gear adapters, so you can relive your '90s gaming experience. TurboGrafx, Neo Geo and Lynx adapters were also announced a long time ago, and will likely arrive one day... one day.

ChatGPT is getting some significant updates that will enable the chatbot to deal with voice commands and image-based queries. Users will be able to have a voice conversation with ChatGPT on Android and iOS and to feed images into it on all platforms. OpenAI is rolling out the features now. They'll be available to Plus and Enterprise users at first, with other folks gaining access to the image-based features later.

You'll need to opt in to voice conversations in the ChatGPT app (go to Settings then New Features) if you'd like to try them out. By tapping the microphone button, you'll be able to choose from five different voices.

OpenAI says the back-and-forth voice conversations are powered by a new text-to-speech model that can generate "human-like audio from just text and a few seconds of sample speech." It created the five voices with the help of professional actors. Going the other way, the company's Whisper speech recognition system converts a user's spoken words into text.

The image-based functions are intriguing too. OpenAI says you can, for instance, show the chatbot a photo of your grill and ask why it won't start, get it to help plan a meal based on a snap of what's in your fridge or prompt it to solve a math problem you take a picture of. As it happens, Microsoft highlighted the Copilot AI's ability to solve math problems in Windows during its Surface event last week.

OpenAI is using GPT-3.5 and GPT-4 to power the image recognition features. To use ChatGPT's image-based functions, tap the photo button (you'll need to tap the plus button first on iOS or Android) to take a snap or choose an existing image on your device. You can ask ChatGPT about multiple photos and use a drawing tool to focus on a specific part of the image.

In a blog post announcing the updates, OpenAI noted the potential for harm. It's possible for bad actors to mimic the voices of public figures (and everyday folks) and perhaps commit fraud. That's why OpenAI is focusing on ChatGPT voice conversations with this technology and working with select partners on other limited use cases (more on that in a moment).

As for images, OpenAI worked with Be My Eyes, a free app that blind and low-vision people can use to help them better understand their surroundings thanks to volunteers who hop into video calls with them. "Users have told us they find it valuable to have general conversations about images that happen to contain people in the background, like if someone appears on TV while you’re trying to figure out your remote control settings," OpenAI said. The company noted that it has also limited how ChatGPT can analyze and make direct statements about people that appear in images, "since ChatGPT is not always accurate and these systems should respect individuals’ privacy." It has published a paper on the safety properties of the image-based functionality, which it calls GPT-4 with vision.

ChatGPT is more effective at understanding English text in images than other languages. OpenAI says the chatbot "performs poorly" in other languages for the time being, particularly when it comes to those that use non-Roman scripts. As such, it suggests that non-English users avoid using ChatGPT to deal with text in images for now.

Meanwhile, Spotify has teamed up with OpenAI to use the voice-based technology for an interesting purpose. The former has announced a pilot of a tool called Voice Translation for podcasters. This can translate podcasts into different languages using the voices of the folks who appear on the show. Spotify says the tool can retain the speech characteristics of the original speaker after converting their voice into other languages.

To start with, Spotify is converting select English-based shows into a few languages. Spanish versions of some Armchair Expert and The Diary of a CEO with Steven Bartlett episodes are available now, with French and German variants to follow.

Getty is officially getting into the AI image business, after banning AI art a year ago. The company's generative AI tool is particularly unique because it's trained on Getty's own database of images, and the resulting content also comes with a royalty-free license. That means you won't have to worry about potential copyright issues down the line. 

While Getty has seemingly been on an anti-AI bent — it's also suing Stable Diffusion maker Stability AI over scraping its data — it's no surprise the company has figured out a way to get into the market. Competing photography services have already announced their own AI solutions: Shutterstock is tapping into Open AI's DALL-E model, while Adobe's Stock is relying on its own Firefly tool as well as accepting some AI generated art.

Getty's Generative AI tool is powered by NVIDIA's Edify AI model from its Picasso cloud service. It'll be interesting to see how those results compare against the improved DALL-E 3 model, but for now it's basically just a way for Getty to dip a toe into the world of AI. As part of this new endeavor, Getty is also launching a "contributor compensation model for AI licensing," Gizmodo reports. Creators will be paid based on the amount of images they've uploaded, together with their existing licensing earnings.

Last year, DJI showed what was possible in a small drone with the Mini 3 Pro by fitting tons of technology and a high-quality camera into a sub-250 gram drone. Following that up was never going to be easy, but now (after numerous leaks) it’s unveiled the Mini 4 Pro with a long list of new features.

Aside from one improvement, the camera is largely the same. However, it has new omnidirectional obstacle sensors that eliminate the blind spots on the Mini 3 Pro. It also comes with a new feature called ActiveTrack 360 that lets you program camera moves when tracking a subject.

Small drones are the best way to track fast-paced action, as they’re maneuverable and less prone to damage when crashing. With all the improvements, the Mini 4 Pro is better and safer at that than its predecessor — at least on paper. Now, let’s find out if it lives up to that in the real world.

Design

The Mini 4 Pro is still under 250 grams so it can be flown without a license or registration in many regions, but it has subtle design changes all around. The four forward- and rear-facing sensors are now placed so that they can see to the sides as well, and the body is a bit more streamlined. It has larger cooling vents, slightly smaller rear arms and new landing feet at the front. The camera/gimbal shield is smaller and easier to put on, and it has a new guard that protects the propellers when it’s stored.

As before, the gimbal tilts up 60 degrees and down 90, and the camera flips 90 degrees to give you full vertical resolution for social media. At the rear is a microSD port, and the Mini 4 Pro has 2GB of internal storage for emergencies.

The Mini 4 Pro comes with one of two controllers, the basic RC-N2 that requires a smartphone, and the RC2 with a built-in screen. Since it uses DJI’s new Ocusync 4 transmission, first introduced with the Air 3, it only works with the new controllers and not the older models — for now, anyway.

You can buy it with a $55 ND filter set for sunny days, which I’d recommend if you can afford it. DJI also offers a wide-angle 18mm equivalent lens attachment ($40), but it has significant barrel distortion and can cause focus issues.

The drone also supports DJI’s Lightcut, an editing app that lets you generate quick videos for social media. As DJI says, it allows “one tap generation of captivating videos by merging ActiveTrack, MasterShots, and QuickShots footage,” while automating sound effects and more. It also works wirelessly, so there’s no need to download footage to your smartphone.

Performance

Given its small size and maneuverability (and the same sensor as DJI's Action 4), you can think of the Mini 4 Pro as a flying action camera. The light weight (and low price compared to, say, a Mavic 3 Pro) also makes crashes less consequential.

Maximum speed is a decently fast 35 MPH in sport mode, or 26 MPH in regular operation. It can handle winds up to 24 MPH, an impressive figure for a sub 250 gram drone. In operation, it can look like it’s being buffeted fairly hard by the wind, but you wouldn’t know it from the footage thanks to DJI’s gimbal and stabilization technology.

Steve Dent for Engadget

The Mini 3 Pro was effectively blind on the sides, but the Mini 4 Pro offers protection all around like the Mavic 3 Pro and Air 3 — thanks to four new omnidirectional sensors on top and two on bottom (along with a time-of-flight sensor). It also uses DJI’s APAS, which offers automatic braking and obstacle bypass for extra security.

If you’re spending $760+ on a drone, you may not want to test the obstacle detection limits. That's my job, so I had it follow me while I walked and biked among trees and other potential snags. I did have a few crashes, but here’s what I learned on how to avoid them.

The sensors are visual, so they don’t work in dim light. And dense forest with fine branches is a no go — the omni sensors can miss those, but the propellers won’t. Finally, the Mini 4 Pro detects obstacles best when traveling forward, less so when going sideways, and worst of all when flying backwards.

It did work around well-spaced trees with thick branches and plenty of leaves, near buildings and generally around well-defined obstacles. It was able to maneuver around those, choose decent routes and reacquire subjects if they disappeared. That helped me capture some nice action footage, though one should always remain wary of accidents.

Steve Dent for Engadget

ActiveTrack 360 adds camera moves to the usual subject tracking to create dramatic shots. It looks confusing at first, but the idea is pretty simple. You use the so-called steering wheel to “draw” a route on concentric circles, and the drone will follow it, ducking any obstacles it encounters.

You can change parameters including the inner and outer radius, inner and outer height, camera speed and ground proximity. That makes it possible to get a wide variety of shots. The tricky part was figuring out which side the drone considered to be forward and backward — DJI should work on this to make things clearer.

If you plan carefully you can get some gorgeous, swooping ActiveTrack shots. The usual obstacle caveats apply, though, and it also adds complexity — because you have to figure out where the drone is going to be when you arrive at your end point. With all that, it’s best to practice in an open area before trying it in a complex environment.

Of course, the Mini 4 Pro still has DJI’s automatic modes aimed at social media users, like MasterShots, QuickShots and Panorama along with Hyperlapse. It even includes the Waypoint feature from the Mavic 3 Pro, which lets you pre-program complex drone moves and repeat them — a sophisticated feature for a small drone.

Steve Dent for Engadget

For the latter feature, you launch the drone and select the Waypoint function, fly to a spot of interest and set the correct camera angle. Once there, you tap “+” on the screen (or hit the C1 button the RC2 remote) to program a waypoint. Repeat that process through all your points of interest, and once you’re done, you can play back the sequence. The drone will smoothly fly to each point the same way each time, so you use it to show a scene during the day and then later at night, for instance.

The Mini 4 Pro uses Ocusync 4 transmission first seen on the Air 3 that sends 1080/60p video up to 20 km, compared to 1080/30p over 12 km with Ocusync 3. Those distances are lower here in Europe because of transmitter power rules.

In use, it provides a noticeable improvement in connectivity, with fewer dropouts and much greater range — even if the drone goes behind obstacles. The change is very noticeable here in Europe with smoother video and dropouts now very rare. It should be even better in the US, where you could send the Mini 4 Pro on a pretty long trip.

Steve Dent for Engadget

The standard 2,590 mAh Intelligent Flight Battery has a bit more capacity than the Mini 3 Pro’s 2,453 mAh cell, but range remains the same at 34 minutes. In real-world flying, we saw about 25 minutes before hearing the return-to-home warning, so plan accordingly.

If you have a Mini 3 Pro, its cells appear to be compatible with the new drone, so that’s a good thing if you already own that model. In the US, you can get the Plus batteries that provide up to 45 minutes of range, but local rules block their use in Europe.

The RC-2 first seen on the Air 3 is DJI’s third screen controller after the RC and the RC Pro. It’s significantly better than the RC, with a brighter screen, better feel and more precise controls. The other option is the non-screen RC-N2 (requiring a smartphone), which is similar to the RC-N1 model but with O4 compatibility.

Camera

Steve Dent for Engadget

The Mini 4 Pro’s camera has the same 1/1.3 dual ISO sensor as its predecessor, using an identical 24mm equivalent lens with a fixed f/1.7 aperture. That’s a pretty sizable sensor for such a small drone, just a bit smaller than the 1-inch sensor on the Mavic Air 2S.

The difference is that it now supports 4K slow-mo at up to 100 fps, or 1080p at 200 fps. The motion is embeded 30fps, but it’s still a nice feature for wildlife, crashing waves and more. That’s on top of 4K at up to 60fps and 1080p at 120fps. It has a two times digital zoom for 4K and four times at 1080p, with a slight loss in sharpness.

There’s also support for DJI’s D-LogM, which boosts dynamic range and gives you more flexibility in post. DJI has a LUT that makes it easy to convert it to regular video — but some editing is required for best results. You can also shoot in DJI’s HLG mode, which again boosts dynamic range. You can see the results right away on an HDR TV, but you’ll need to do an HLG to REC.709 color-space transform to use it with regular video. Both support 10-bit 4:2:0 capture for improved fidelity and reduced banding.

Quality is about the same as a really good smartphone — but not on par with a mirrorless camera or DJI’s pro-level Mavic 3. Video is sharp with accurate colors. The automatic mode delivers nice video, though it sometimes over- and underexposes on sunny or dark days. You can change exposure compensation, but nothing else. Luckily, a fully manual pro mode is available for better control of color balance, LOG, HLG, shutter, ISO and more

The Mini 4 Pro can shoot sharp 48-megapixel images or combine four pixels into one for 12-megapixel images with improved night sensitivity. You can easily fix over- or under-exposed photos if you use the RAW DNG format.

Low-light sensitivity is good but not great —- better than, say, a GoPro 12. Shooting at twilight, video was less clear than a similar scene shot with the Mavic 3 Pro. The drone also offers a “night” mode that effectively boosts dynamic range, making dimly-lit scenes pop better.

In all, image quality isn’t perfect, but remember that this is a $1,000 lightweight drone. It beats all other models in that category, and it’s better than many heavier drones, too.

Wrap-up

Steve Dent for Engadget

Once again, DJI’s Mini 4 Pro sets a benchmark for small drones. It has multiple new useful features, including updated obstacle detection, ActiveTrack 360, O4 transmission and Waypoints. All of those make it a solid budget choice for action sports, events, aerial photography, industrial applications and more.

Its main competition is the Autel Evo Nano Plus, currently on sale for $580. That model has a similar 50-megapixel 1/1.27-inch camera sensor, three-way obstacle avoidance, subject tracking, and more. However, it’s limited to 4K 30p and doesn’t offer a remote with a screen. If you have a bit more to spend, DJI’s Air 3 offers more stability and an extra tele camera.

All that said, the Mini 4 Pro isn’t cheap for a budget drone. It’s priced at $759 for the drone with a battery and RC-N2 controller, $959 with the RC2 controller and $1,099 for the Fly More kit with three batteries and a charger, the RC2, a carrying case and extra props. Still, if you’re in the market for a drone in that price range, nothing else can really touch it.

Hey big spender. Tinder Select, the dating app’s most exclusive tier, is rolling out now. It will cost love seekers $500 per month (or $6,000 annually — no bulk discounts) for features like exclusive search and matching.

The company has only offered Tinder Select to the less than one percent of users it considers “extremely active” — does anyone want that label? Tinder told Bloomberg it’ll open applications for Tinder Select on a rolling basis, but it didn’t say exactly when. Tinder’s exclusive membership was originally hinted at all the way back in 2019.

The owners of Tinder, Match Group, have dabbled in exclusive dating apps before, like The League, which it bought in 2022, so it’s not too much of a shock to see Tinder also get reframed for the lonely rich. Is this worse than paying for verification when you have less than 1,000 followers on other social media networks? Yes. Yes, it is.

— Mat Smith

The biggest stories you might have missed

What the Elon Musk biography revealed about his tumultuous Twitter takeover

Drop BMR1 PC speaker review: Not bad, but not amazing

The best October Amazon Prime Day early access deals for 2023

Hitting the Books: Beware the Tech Bro who comes bearing gifts

The Morning After: Microsoft’s bad week, and Alexa gets an attitude

Last week’s biggest news meets Engadget’s lens.

Engadget

Our short-but-sweet YouTube edition of this week’s news covers includes Microsoft’s rough, rough week, a sassier Alexa from Amazon and whether the iPhone 15 Pro is worth the extra bucks. Also: viewers take umbrage at my ‘fake’ glasses. Which are not fake.

Watch here.

Sony ZV-E1 camera review

The best vlogging camera, by a big margin.

Engadget

I’ve been waiting for this. Sony fully embraced amateur / semi-pro content creators back in 2020, with the launch of the ZV1 camera. It has since added no less than four models to its ZV lineup, and this is the latest: the 12-megapixel full-frame ZV-E1. It uses the same sensor as the $3,500 A7S III, a video-focused camera — and a low-light marvel. However, the ZV-E1 costs $1,300 less. While Sony has cut some minor corners, it combines outstanding video features and AI tricks, and I might have to start saving for one. 

Check out the full review.

Samsung leaks its next family of smartphones, earbuds and tablets

Don’t get too excited. It’s the Fan Edition ones.

Samsung

Eagle-eyed visitors to Samsung’s Argentinian website — I visit it weekly — have spotted something a little unexpected: a product page for new Galaxy Buds FE earbuds, along with images of a Galaxy S23 FE smartphone and Galaxy Tab S9 FE tablet. Samsung’s Fan Edition devices have proven popular, packing in solid features for a more reasonable price than Samsung’s flagship models.

The company hasn’t let slip any specs for the phone and tablet yet. However, there are some details on the Galaxy Buds FE, Samsung’s first Fan Edition earbuds. They’re slated to have a single 12mm driver, three microphones in each earbud to bolster active noise cancellation and a three-way speaker.

Continue reading.

The best foldable phones for 2023

Are flip phones back?

Foldables have come a long way since the original Galaxy Fold went on sale back in 2019. They’re smaller, they’re tougher and, while they still aren’t a great option for people on a budget, they’re now more affordable too. (Kind of?) We walk through the crucial specs, durability concerns and our favorite picks.

Continue reading.

The Engadget Podcast

iPhone 15 Pro reviews, and Microsoft picks AI over Surface.

This week, Cherlynn chats about her experience reviewing the iPhone 15 Pro and Apple Watch Series 9. Does a 5X camera zoom make much of a difference? Meanwhile, Microsoft is basically consolidating all of the Copilot products it’s already announced for Edge, MS 365 and Windows, but maybe this will be less confusing in the long run?

Continue reading.

Amazon is investing up to $4 billion in OpenAI rival Anthropic as a way to provide advanced deep learning and other services to its Amazon Web Service (AWS) customers, the company wrote in a press release. In return, AWS becomes Anthropic's "primary cloud provider" to train and deploy its future foundation models. It's the second large investment in the company, founded by former OpenAI executives, following Google's $400 million partnership with the firm. 

The e-commerce company will start with a $1.25 billion investment to gain a minority stake in Anthropic, with an option to boost that to a total of $4 billion. Along with Google and Amazon, Anthropic also counts Salesforce, Zoom, Spark Capital and others as backers. Notably, Anthropic's deal with Google didn't require it to buy cloud services from the search giant. 

Anthropic recently unveiled its first consumer-facing chatbot Claude 2, accessible by subscription much like OpenAI's ChatGPT. The Claude “Constitutional AI” system is guided by 10 “foundational” principals of fairness and autonomy and is supposed to be harder to trick than other AI. Anthropic is also working on a chatbot it calls "Claude-Next" that's supposed to be ten times more powerful than any current AI, according to TechCrunch. 

The startup touts itself as an advocate for responsible AI deployment, and recently formed an AI safety group with Google, Microsoft and Open AI. It has been with AWS since 2021. "Claude excels at a wide range of tasks, from sophisticated dialogue and creative content generation to complex reasoning and detailed instruction, while maintaining a high degree of reliability and predictability," according to Amazon.  

Instead of training their own models, AWS customers will be able to use Anthropic's AI models via Amazon's Bedrock, a service designed specifically for AI development. Amazon Cloud also offers its own AI applications, and with the new partnership, is hoping to position itself as a key player in the field.

Microsoft-backed OpenAI is largely considered to be the leader in AI and chatbot tech, thanks to its ultra-popular ChatGPT chatbot and DALL-E image generation service. Use of AI in business continues to grow exponentially, despite concerns over the legality and ethics of AI-appropriated content — it was considered to be a strong sticking point in the WGA writer's strike, for example.