techcrunch
Author page description
August 19
- Ikea Expands Its Inventory Drone Fleet➀ Ikea is expanding its use of drones for inventory management, partnering with Verity to deploy drones in its stores for real-time stock monitoring. ➁ The drones are designed to navigate store aisles autonomously, scanning barcodes and identifying misplaced items. ➂ This technology aims to improve inventory accuracy and efficiency, reducing the need for manual stock checks.
August 18
- Google Challenges OpenAI with Gemini Live➀ Google unveiled its new lineup of Pixel 9 phones, including the $1,799 Pixel 9 Pro, at its Made by Google event. ➁ The company also announced Gemini Live, a direct competitor to OpenAI's advanced AI models. ➂ Google is expanding its AI-powered search summaries to six more countries: India, Brazil, Japan, the U.K., Indonesia, and Mexico.
June 15
- GM’s BrightDrop makes first commercial EVs delivery outside USGeneral Motors的电动送货车子公司BrightDrop已经完成了在美国之外的第一次商业交付。该公司于本周向FedEx Express Canada交付了50辆BrightDrop Zevo 600电动货车,这是一份包含2,500辆Zevo EV的大订单的一部分。这些电动车将为FedEx Canada的快递服务提供支持。这些电动车可用于最后一公里的快递送货,据估计可行驶250英里。FedEx的目标是到2040年将其整个车队转变为全电动车辆。BrightDrop的Zevo 600型号可用于最后一英里的送货,估计行驶里程为250英里。FedEx已经收到约400辆Zevos,目前正在南加州运营。此外,GM还在制造一种名为Zevo 400的更小型号。 全文: TechCrunch Open Navigation TechCrunch Rebecca Bellan BrightDrop, the electric delivery van subsidiary of General Motors, has made its first delivery outside of the U.S. The company delivered 50 BrightDrop Zevo 600 electric vans to FedEx Express Canada on Wednesday, part of a larger order of 2,500 Zevo EVs. The introduction of BrightDrop’s vans is in line with FedEx’s goal to shift its entire fleet to all-electric vehicles by 2040. BrightDrop’s Zevo 600 model is designed for last-mile deliveries and has an estimated range of 250 miles. FedEx has already received about 400 Zevos which are currently in operation in Southern California. FedEx Canada’s first fleet of BrightDrop vehicles will be seen delivering parcels on the streets of Toronto, Montreal and Surrey, according to FedEx. The company said it is working to install charging infrastructure across its Canadian facilities. The vehicles were assembled at GM’s CAMI Assembly plant in Ontario, which began production in December 2022. GM invested around $750 million ($1 billion CAD) into the plant, which is expected to produce 50,000 Zevo vans annually by 2025. Aside from the Zevo 600, GM is also building a smaller version, called the Zevo 400 . FedEx is one of a handful of BrightDrop’s big commercial partners. The company has received around 30,000 reservations and letters of intent from customers including Walmart , Hertz, Verizon and most recently Ryder . Rivian, one of BrightDrop’s e-delivery van competitors, started rolling out its vans in summer 2022 and has delivered around 3,000 units to Amazon, according to the e-commerce giant’s blog . Amazon hopes to have 100,000 Rivi
- Radio station gets part-time AI DJ based on its midday host美国俄勒冈州波特兰的电台Live 95.5宣布,该电台将每天从上午10点到下午3点向听众播放中午节目主持人Ashley Elzinga的克隆声音,即“AI Ashley”。这是通过使用Futuri Media的“RadioGPT”实现的,该工具基于ChatGPT生成基于热门新闻的脚本,并以合成语音读出。Alpha Media的执行副总裁Phil Becker表示,AI Ashley不会完全取代传统Ashley,而是在某些片段中使用传统Ashley,在其他片段中使用AI Ashley。Alpha Media管理着美国200多个电台。Futuri的技术扫描超过250,000个新闻来源以及Facebook、Twitter和Instagram上的故事,识别本地热门话题,并使用GPT-3创建脚本。在AI语音向当地听众传递信息之前,一个人工模式团队会审查和编辑内容。 全文: TechCrunch Open Navigation TechCrunch Lauren Forristal Live 95.5, a radio station in Portland, Oregon, announced Tuesday that its midday host Ashley Elzinga will broadcast a cloned version of her voice — aka “AI Ashley” — to listeners every day from 10 a.m. to 3 p.m. They’ll be using Futuri Media’s “RadioGPT ,” an AI-powered tool that uses ChatGPT to generate a script based on trending news and reads it with a synthetic voice. The radio station noted that AI Ashley isn’t fully replacing “traditional Ashley,” which has been a huge concern for a lot of local DJs. Radio stations have cut down their broadcasts in recent years due to rising costs or in a shift to embrace tech like AI. In 2020, iHeart Media restructured its organization, laying off hundreds of people nationwide because it wanted to invest in artificial intelligence. Live 95.5’s parent company, Alpha Media, assured us that Elzinga would not be losing her job and would still be paid the same salary. “It’s a hybrid situation where we’ll have traditional Ashley on during some segments, and we’ll have AI Ashley on during other segments,” Phil Becker, Alpha Media EVP of Content, explained to TechCrunch. “In an instance where AI Ashley would be broadcasting, the traditional Ashley might be doing something in the community, managing social posts or working on digital assets or the other elements that come with the job.” Becker also noted that Alpha Media isn’t using RadioGPT to save costs. It’s meant to be an efficient tool for radio hosts to have in their toolset. Listeners may be wary of AI DJs, since they have tuned int
- EV startup Telo bets America is ready for a dreamy little pickupTelo Trucks, a young EV startup, is betting on America's interest in a petite pickup truck with the "footprint of a two-door Mini Cooper" and the same interior and bed space as a Toyota Tacoma. CEO Jason Marks says the company is targeting city dwellers who want a cargo utility vehicle, such as gardeners, surfers, snowboarders and hikers, but can't have a big truck because it's too big for the city. Telo has received more than 500 preorders within the first 12 hours of the launch, according to a company representative. Telo has raised $1.4m so far at a $10m post-money valuation and is working towards getting its first prototype on the road by year's end. 全文: TechCrunch Open Navigation TechCrunch Harri Weber Most trucks today are headed in the wrong direction, metaphorically speaking. They’re far too big and heavy, with shrinking beds and expanding cabins that reflect their turn from a classic workhorse into a status symbol-meets-family car. This is my take, but I’m not just here to blab about aesthetics and my weird love of teeny-tiny cars. Huge vehicles are uniquely deadly for pedestrians and cyclists and counterproductive to decarbonization work. They demand more raw materials and ultimately bigger batteries , and they stir fears that compact cars can’t hack it alongside them on roads today. In other words, they’re bad for people and the planet. The switch to electric vehicles offers an opportunity to shake things up, sizewise. Yet in the U.S., most automakers won’t risk it, and there’s a financial basis for that. Trucks like the F-150 and Silverado are more popular than ever , giving little incentive for companies to carve out a new path. Still, a young startup called Telo Trucks is taking an alternate route anyway. It’s betting that lots of Americans actually want a petite pickup — one with the “footprint of a two-door Mini Cooper” and the “same interior and bed space as a Toyota Tacoma,” Telo CEO Jason Marks told TechCrunch. Do you believe him? Marks ar
- TrueCar lays off 102 employees, taps new CEO amid restructure汽车价格研究平台TrueCar近日宣布进行重组,裁员102人,占员工总数的24%。公司总裁兼CEO Michael Darrow将卸任执行职务,由之前担任TrueCar首席运营官和首席财务官的Jantoon Reigersman接任。TrueCar表示,重组将使公司的费用结构更好地与收入基础相匹配,使其成为一个更加敏捷、高效的公司。除了股票补偿之外,TrueCar预计其年化费用将减少超过2,000万美元。重组将导致大约700万美元的非经常性现金支出,这笔支出将在2023年第二季度和第三季度出现。TrueCar的第一季度收入下降,成本上涨,公司的净亏损达1950万美元,高于2022年第一季度的1240万美元。 全文: TechCrunch Open Navigation TechCrunch Rebecca Bellan TrueCar , a platform where people can research new and used car prices and find local dealer savings, is laying off 24% of staff due to a restructure, the company said Wednesday. About 102 people will be impacted by the layoff. President and CEO Michael Darrow is stepping aside both from his executive roles — to be replaced by Jantoon Reigersman — and his position on the board. Reigersman previously served as TrueCar’s chief operating officer and chief financial officer. TrueCar’s stock jumped nearly 13% in after-hours trading on the news. “The restructuring announced today better aligns our cost structure with our revenue base and is designed to make TrueCar a nimbler, more efficient company,” said Barbara Carbone , incoming chair of the board, in a statement. “We made this difficult decision after an extensive review and believe that it is necessary to enable TrueCar to achieve its strategic priorities and create long-term shareholder value.” TrueCar says the restructure should result in nonrecurring cash payments of about $7 million, which will show up in the second and third quarters of 2023, and an annualized reduction in expensive exclusive of stock-based compensation of over $20 million. As of May 31, 2023, TrueCar had cash and cash equivalents of $146.5 million. The company expects that number to drop below $125 million in the near term, but says it continues to anticipate break-even or positive adjusted EBITDA and double digit year-over-year revenue growth in the fourth quarter of 2023. TrueCar’s first-quarter earnings also show a dip in revenue and increase in cost. The company closed Q1
- Fernride nudges yard trucks towards full autonomy with $31M in new funds全文: TechCrunch Open Navigation TechCrunch Rebecca Bellan Fernride has figured out a way to launch a commercially viable autonomous, electric truck business today. The secret? Not driving on public roads. Most AV trucking companies are going after the moonshot: using self-driving freight trucks — typically 18-wheelers — to haul goods over long distances. Startups like Waymo , TuSimple , Kodiak Robotics and Aurora have all had commercial pilot projects and are actively testing on public highways in Texas, Arizona and other southern states. And none of them are operating without a human safety operator in the front seat; most are years away from commercialization. Fernride, a Munich-based startup founded in 2019, is in a different kind of trucks business. It’s focused on the yard trucks used to move trailers and containers around ports, terminals and distribution facilities. “We wanted to do things completely differently,” said Hendrik Kramer, CEO and co-founder of Fernride. “First, we wanted to focus on the use case that works today, specifically yards or geofenced areas on private sites where we can launch a product now. Then when we have launched that first product and built customer relationships, we can scale to the open roads.” Fernride’s trucks have Level 4 autonomous capabilities, which means they can drive themselves without human intervention under certain conditions. There is a remote driver who monitors the self-driving trucks and can assist the vehicles if needed. Fernride calls this “human-assisted autonomy” — a combination of full Level 4 driving and tele-operated driving that allows one remote driver to control four trucks at a time. This approach, Kramer says, guarantees 100% operational availability of the system from the get-go. Fernride is currently working with a fleet of six trucks and plans to scale to 20 by the end of 2023. The startup locked in a partnership with Terberg, a Dutch yard truck manufacturer, to begin series production on
- Why every EV charging network combined can’t compete with Tesla全文: TechCrunch Open Navigation TechCrunch Tim De Chant It’s no secret that the charging infrastructure for electric vehicles generally sucks. There are exceptions, of course: Tesla has it pretty well figured out , and some highly trafficked corridors are well covered. But overall, the state of fast charging, which can replenish usable amounts of range in 30 minutes or less, isn’t great. There are plenty of reasons why. Most chargers are located in massive parking lots, usually in a forgotten corner with little in the way of amenities for EV drivers. The equipment itself is notoriously unreliable, with one study suggesting that about a quarter of all Combined Charging System-compatible (CCS) stalls in the Bay Area are out of service at any given time. While charging speeds are increasing, most chargers aren’t nearly as fast as they need to be. Some of those problems are easier to swallow than others. But one not mentioned above is a deal breaker: the dearth of available chargers. When they’re in short supply, EV drivers either struggle to find a spot or have to wait in line, sometimes for a while. It’s widely accepted that Tesla’s Supercharger network is the best. It’s broadly distributed, the chargers are generally reliable, and most importantly, numerous. That’s in part because Tesla’s fleet is the largest fully electric fleet in the U.S., with around 1.6 million vehicles on the road. It makes sense that their network is also the largest, with 17,551 stalls that charge at 120 kW or greater, according to data from the Department of Energy and Supercharger.info .
- Mechanical Turk workers are using AI to automate being human全文: TechCrunch Open Navigation TechCrunch Devin Coldewey File this one under inevitable, but hilarious. Mechanical Turk is a service that from its earliest days seemed to invite shenanigans, and indeed researchers show that nearly half of its “turkers” appear to be using AI to do tasks that were specifically intended to be done by humans because AI couldn’t. We’ve closed the loop on this one, great job everybody! Amazon’s Mechanical Turk let users divide simple tasks into any number of small sub-tasks that take only a few seconds to do, and which pay pennies — but dedicated piecemeal workers would perform thousands and thereby earn a modest but reliable wage. It was, as Jeff Bezos memorably put it back then, “artificial artificial intelligence.” These were usually tasks that were then difficult to automate — like a CAPTCHA, or identifying the sentiment of a sentence, or a simple “draw a circle around the cat in this image,” things that people could do quickly and reliably. It was used liberally by people labeling relatively complex data and researchers aiming to get human evaluations or decisions at scale. It’s named after the famous chess-playing “automaton” that actually used a human hiding in its base to make its plays — Poe wrote a great contemporary takedown of it . Sometimes automation is difficult or impossible, but in such cases you can make a sort of machine out of humanity. One has to be careful about it, but it has proven useful over the years. But a study from researchers at EPFL in Switzerland shows that Mechanical Turk workers are automating their work using large language models like ChatGPT: a snake biting its own tail, or perhaps swallowing itself entirely. The question emerged when they considered using a service like MTurk as a “human in the loop” to improve or fact-check LLM responses, which are basically untrustable: It is tempting to rely on crowdsourcing to validate LLM outputs or to create human gold-standard data for comparison.
- Ask Sophie: How much time and money will we need for an H-1B transfer?一位初创公司的联合创始人向移民律师Sophie Alcorn提问,该公司正在招聘,许多优秀的工程师需要进行H-1B转移,但他还没有进行过这项操作。该联合创始人询问了需要多少时间和资金,是否有替代方案。Sophie建议寻找经验丰富的移民律师,以节省时间和资金,并确定最佳策略和协助公司提交新员工的请愿书。通常情况下,初创公司进行H-1B专业职业签证转移的时间为一个月左右,费用通常不到10000美元。Sophie还解释了转移流程,包括验证公司的FEIN,提交劳动条件申请,根据工作和地理位置确认该公司将支付H-1B转移候选人的工资,并确保没有合格的美国工人填补该职位。 全文: TechCrunch Open Navigation TechCrunch Sophie Alcorn Sophie Alcorn Contributor Share on Twitter Sophie Alcorn is the founder of Alcorn Immigration Law in Silicon Valley and 2019 Global Law Experts Awards’ “Law Firm of the Year in California for Entrepreneur Immigration Services.” She connects people with the businesses and opportunities that expand their lives. More posts by this contributor Ask Sophie: Why is there no movement in the June Visa Bulletin for India EB-3? Ask Sophie: How long until I can travel while waiting for my green card? Here’s another edition of “Ask Sophie,” the advice column that answers immigration-related questions about working at technology companies. “Your questions are vital to the spread of knowledge that allows people all over the world to rise above borders and pursue their dreams,” says Sophie Alcorn , a Silicon Valley immigration attorney. “Whether you’re in people ops, a founder or seeking a job in Silicon Valley, I would love to answer your questions in my next column.” TechCrunch+ members receive access to weekly “Ask Sophie” columns; use promo code ALCORN to purchase a one- or two-year subscription for 50% off. Dear Sophie, My startup is hiring and many excellent engineers need H-1B transfers, but I haven’t done one yet. Approximately how much time and money will we need to set aside for the process? Are there alternatives? — Careful Co-Founder Dear Careful, Congrats on making it to the next stage: hiring! Working with an experienced immigration attorney can help you save time and money in the process so you can onboard your new hires rapidly and continue to build. It’s important that your attorney understands your startup’s v
- Canva introduces new developer tools and $50M investment fund设计平台Canva宣布推出新的开发工具和5000万美元的投资基金。新的软件开发工具包(SDK)和API可以让开发者在Canva平台上构建应用程序,包括添加文本、图像、视频、创建自己的形状和插图等。Canva还发布了五个API,包括上传媒体到Canva库的"Asset",添加数据到Canva的"Data",构建新设计工具的"Design",发送HTTP请求的"Fetch"和认证的"User"。Canva还宣布推出应用程序市场来分发在平台上构建的应用程序,并设立5000万美元的基金,帮助推出一些更具前景的创意。Canva的联合创始人兼首席产品官Cameron Adams表示,Canva拥有1.35亿用户,开发者可以通过平台接触到这些用户,成为设计生态系统的一部分。 全文: TechCrunch Open Navigation TechCrunch Ron Miller Canva has been around since 2013, and has been a tremendously successful design platform, raising over $570 million and garnering 135 million users along the way. With that kind of traction, it makes a ton of sense to let developers tap into that user base and take advantage of the work that the company has already done. Today, at the first Canva Extend developer conference in San Francisco, the company announced a new SDK, more robust APIs and a $50 million fund to help encourage some of the better ideas. Cameron Adams, co-founder and chief product officer at the company, says they’re staging the developer-focused event because it’s a really new phase in what’s going to be possible for developers interacting with Canva. For starters, the company is releasing a software development kit (SDK), which Adams calls “one of the slickest experiences I’ve had developing on a platform. It is super polished; the team has done amazing work on it,” he told TechCrunch. This enables companies to build applications on top of the Canva platform. They’ve also released a slew of new APIs. “So previously, you could build a bit on top of Canva, but with the new APIs we’ve just totally unlocked our platform. So now any developer can start creating an app that allows him or her to truly interact with people’s designs. So now they can add text elements, image elements, video elements, they can create their own shapes and illustrations that end up in people’s designs,” Adams said. The announcement includes five APIs including Asset for uploading media to the Canva library, Data to add data to Canva,
- Sony readies PS5 game streaming for PlayStation Plus Premium membersSony开始测试PlayStation 5游戏流媒体服务,并计划将其作为PlayStation Plus Premium会员的附加福利。此举标志着Sony云游戏策略的演变,此前用户必须下载游戏才能从主机上远程流媒体播放。该公司表示,随着越来越多的游戏在PS5主机上推出,他们期待为PS5游戏添加云游戏功能。此举使Sony的策略更加符合Microsoft的Xbox Game Pass服务所提供的内容,订阅者可以更轻松地享受游戏。此外,Sony还公布了一些新的游戏目录,包括Far Cry 6、Teenage Mutant Ninja Turtles: Shredder's Revenge等游戏。 全文: TechCrunch Open Navigation TechCrunch Brian Heater Sony this morning announced that it has begun testing PlayStation 5 game streaming . The feature is still early days, but it will eventually be available for those subscribed to PlayStation Plus Premium. Sony notes, “Our goal is to add this as an additional benefit to PlayStation Plus Premium as part of our ongoing efforts to enhance the value of PlayStation Plus.” The move marks an evolution for the company’s cloud gaming strategy, which previously required users to download the title in order to stream it remotely from the console itself. “We think it’s important for Premium members to be able to enjoy as many games as possible via cloud streaming,” Sony VP Nick Maguire said in a blog post. “As more games continue to launch on the PS5 console, we look forward to adding cloud streaming capability for PS5 titles in addition to the PS3, PS4 and classic titles that are already available for Premium members to stream.” The long-awaited move brings Sony’s strategy more in line with what Microsoft has long offered through its Xbox Game Pass service. It’s frankly an easier and more friction-free experience for the subscriber, so it makes sense that Sony’s following suit here. Image Credits: Sony The testing also, notably, arrives ahead of Sony’s Project Q device . Last month’s reveal of the game-streaming handheld was greeted with a bit of a mixed reaction only. Certainly removing the need to stream directly from a separate PlayStation console makes the system at least a bit more appealing. Sony has promised more information on the Project Q at some time in the near future. Ditto for timing on the new streaming fun
- Twitter is being evicted from its Boulder office over unpaid rentTwitter is facing eviction from its Boulder office due to unpaid rent, with the landlord drawing on a $968,000 letter of credit to pay rent in lieu of ordinary payments. The amount ran out in March, and Twitter has not paid since. A judge has ordered the sheriff to assist in the eviction before the end of July. As many as 300 employees once worked in Twitter’s Boulder offices, but it is probably less than half of that now due to layoffs, other firings, and resignations. The news follows reports of unpaid bills and disarray since Elon Musk took over the company. 全文: TechCrunch Open Navigation TechCrunch Devin Coldewey Twitter owes three months’ rent to its Boulder landlord, and a judge has signed off on evicting the tech giant from its office there, court documents show. Since its takeover by Elon Musk , Twitter’s business has more or less fallen into disarray, and there have been numerous reports of unpaid bills . While a contractor going unpaid during a rocky transition is unfortunate but not uncommon, ceasing to pay rent altogether for months suggests Twitter’s operations may be farther gone than anyone expected. According to court documents and reporting by the Denver Business Journal , Lot 2 SBO LLC, the Chicago-based landlord that owns Twitter’s offices at 3401 Bluff St in Boulder, was provided a $968,000 letter of credit back in February of 2020, which it has been drawing on to pay the rent in lieu of ordinary payments (the details of this arrangement are somewhat obscure). That amount ran out in March, and the company has not paid since; if we assume rent was paid regularly from that sum, that places it at around $27,000 per month. In May the landlord took it to court, and on May 31 the judge issued an order that the sheriff should assist in the eviction of Twitter within the next 49 days — i.e. before the end of July. The case number is 2023CV30342 in Boulder District Court. As many as 300 employees once worked in Twitter’s Boulder offices, but betw
- Former SEC chair Jay Clayton feels ‘vast majority’ of crypto tokens are securitiesFormer SEC Chairman Jay Clayton believes that many cryptocurrencies could be defined as securities, stating that "many, if not the vast majority, of the tokens that were sold for cash would fall within the definition of a security in America". However, Clayton also noted that something once labeled as a security "might not always be a security" based on present utility versus future utility. He used the example of Broadway show tickets, stating that if someone bought 1,000 tickets and told their friends and family they could resell them for a profit, it would be a security, but if they just bought the ticket 10 years later, it would not be. Clayton believes there should be a digitised market for non-securities trading. 全文: TechCrunch Open Navigation TechCrunch But there's hope: Something that is once labeled as a security, 'might not always be a security.' Jacquelyn Melinek Former SEC Chairman Jay Clayton has reiterated his position that many cryptocurrencies could be defined as securities, even as the crypto industry continues to combat the U.S. Securities and Exchange Commission over the regulator’s prohibitive stance towards the industry. “I’ve said this for a long time: I think the market has evolved, but many, if not the vast majority, of the tokens that were sold for cash would fall within the definition of a security in America,” Clayton, now a senior policy advisor and counsel at Sullivan & Cromwell LLP, said at the R3 CordaDay conference on Wednesday. The definition of a security is “intentionally broad and flexible,” Clayton noted. But, he added that there’s a chance that something once labeled a security, “might not always be a security.” So what could cause that shift? Present utility versus future utility, Clayton said. Clayton pointed to broadway show tickets as an example: If someone bought 1,000 tickets for $10 and told their friends and family they would be able to resell those tickets for $100 or $1,000, then it’s a security, he said. “Bu
- Synthesia secures $90M for AI that generates custom avatarsSynthesia, a startup that uses artificial intelligence (AI) to create synthetic videos for advertising and other use cases, has raised $90m in a series C funding round led by Accel, with a strategic investment from Nvidia and participation from Kleiner Perkins, GV, Firstmark Capital and MMC. The company uses AI to create instructional videos featuring stock or custom AI avatars. Synthesia's AI is trained on real actors, who are paid per video generated with their image and voice. Synthesia has over 50,000 customers and a year-on-year user growth rate of 456%, with over 15 million videos generated on the platform to date. The investment will be used to make Synthesia's avatars more expressive and the platform faster and "more collaborative". 全文: TechCrunch Open Navigation TechCrunch Kyle Wiggers Once the pandemic normalized virtual meetups, the concept of “personalized AI” began to gain steam. Startups creating “AI-driven” avatars — realistic-looking characters with synthetic voices that star in pre-recorded or live videos — have raised hundreds of millions of dollars in venture capital over the past few years. As such avatars improve, they promise to deliver more personalized digital marketing and training experiences while reducing the costs typically associated with video production. That’s at least the sales pitch from Synthesia , one of the startups using AI to create synthetic videos for advertising and other use cases. The company yesterday announced that it raised $90 million in a Series C round led by Accel with a strategic investment from Nvidia and participation from Kleiner Perkins, GV, Firstmark Capital and MMC. The round brings Synthesia’s total raised to date to $156.6 million, and it values the startup at $1 billion post-money (up from $300 million in December 2021). Co-founder and CEO Victor Riparbelli tells TechCrunch that Synthesia is a “sustainable business” and wasn’t looking for an additional investment, but that Accel and Nvidia approac
- Google Lens can now search for skin conditions谷歌的计算机视觉应用程序Google Lens现在可以搜索皮肤状况。用户可以通过Lens上传照片并搜索与自己皮肤相似的症状,例如痣和皮疹。此外,Lens还可以搜索其他身体疾病的视觉匹配项,例如唇部肿块、指甲上的线条或脱发。虽然这一功能比谷歌2021年推出的诊断皮肤、头发和指甲状况的AI应用程序少了一步,但对于决定是否寻求医疗治疗或购买非处方药治疗的人来说仍然可能很有用。此外,Lens还与谷歌的AI聊天机器人Bard集成,用户可以在他们的Bard提示中包含图像,Lens将在幕后帮助Bard理解所显示的内容。 全文: TechCrunch Open Navigation TechCrunch Kyle Wiggers Google’s enhancing Google Lens, its computer vision-powered app that brings up information related to the objects it identifies, with new features. Starting today, Lens can surface skin conditions similar to what you might see on your own skin, like moles and rashes. Uploading a picture or photo through Lens will kick off a search for visual matches, which will also work for other physical maladies that you might not be sure how to describe with words (like a bump on the lip, a line on nails or hair loss). It’s a step short of the AI-driven app Google launched in 2021 to diagnose skin, hair and nail conditions. That app, which debuted first in the E.U., faced barriers to entry in the U.S., where it would have had to have been approved by the Food and Drug Administration. (Google declined to seek approval.) Image Credits: Google Still, the Lens feature might be useful for folks deciding whether to seek medical attention or over-the-counter treatments. Elsewhere, as previously announced at I/O, Lens is integrating with Bard, Google’s AI-powered chatbot experience. Users will be able to include images in their Bard prompts and Lens will work behind the scenes to help Bard make sense of what’s being shown. For example, shown a photo of shoes and asked what they’re called, Bard — informed by Lens’ analysis — will come back with an answer. It’s the latest update to Bard, Google’s answer to ChatGPT, as Google invests an increasing amount of resources into generative AI technologies. Just last week, Google introduced a capability that allows Bard to write, execute and test its own code in the background — improving its ability to program and solve complex ma
- EVs are going backwards电动汽车(EV)面临着一个难题,随着电池容量的增加,车辆的重量也会增加,从而影响其燃油效率。尽管技术上车辆电池充电时会变重,但差异非常小。重量是影响车辆燃油效率的最重要因素。例如,1984年大众兔只有90马力,但平均油耗为每加仑34英里。相比之下,2009年版的大众兔几乎有两倍的马力和各种技术优势,但平均油耗仅为每加仑24.6英里。这是因为它比旧款车型重近1000磅。因此,更小和更环保的车辆是未来的趋势。而目前,GM取消Bolt系列是朝着错误方向迈出的巨大一步。 全文: TechCrunch Open Navigation TechCrunch Smaller and greener is the future Haje Jan Kamps When you send a rocket into space, you run into a paradox: Around 90% of the weight of the rocket is the fuel that it needs to lift away from Earth. As the rocket accelerates, three things happen: The speed increases, the rocket gets further away from Earth (and from the gravitational pull of the planet), and as the fuel burns off, the rocket weighs less and so can accelerate faster. We face a similar problem in the world of EVs, with an obvious exception. Technically , car batteries get heavier as they are charged , but the difference is so small that you couldn’t measure it even with the most precise car scale. However, that doesn’t matter, because when you add more batteries to a car, it must always carry all those batteries. That’s a problem that’s growing bigger as cars themselves get bigger and heavier. A 1984 Volkswagen Rabbit only had 90 horsepower, but its average fuel economy was 34 miles per gallon . The 2009 version of that car had almost twice the horsepower and all sorts of technology advantages, but its average fuel consumption is a paltry 24.6 miles per gallon . Sure, the engines got more efficient and it has a bunch of new comfort features. It’s safer, too, but it weighs almost 1,000 lbs more. Weight, more than almost anything else, affects the fuel economy of a vehicle. It turns out that batteries are quite heavy. The battery pack on a Tesla Model 3 weighs more than 450kg (1,000 lbs). The Rivian truck’s battery weighs 800kg (1,755 lbs). And if you buy an EV Hummer, the battery you’re dragging around weighs more than a 2009 Volkswagen Rabbit at 1,300kg (2,800 lbs). There’s a better way. American drivers aren’t going to
- Google’s new generative AI lets you preview clothes on different modelsGoogle has launched a new generative AI-powered shopping feature that allows users to preview clothing items on a diverse range of real-life fashion models. The virtual try-on tool, which will be available for women's tops from brands such as Anthropologie, Everlane, H&M and LOFT, uses a diffusion-based model that aims to predict how clothing will drape, fold, cling, stretch and form wrinkles and shadows on models in different poses. Google trained the model using pairs of images of people wearing garments in two unique poses, and repeated the process using random image pairs. 全文: TechCrunch Open Navigation TechCrunch Kyle Wiggers Google, ever eager to lean into generative AI, is launching a new shopping feature that shows clothes on a lineup of real-life fashion models. A part of a wide range of updates to Google Shopping rolling out in the coming weeks, Google’s virtual try-on tool for apparel takes an image of clothing and attempts to predict how it would drape, fold, cling, stretch and form wrinkles and shadows on a set of real models in different poses. Virtual try-on is powered by a new diffusion-based model Google developed internally. Diffusion models — which include the text-to-art generators Stable Diffusion and DALL-E 2 — learn to gradually subtract noise from a starting image made entirely of noise, moving it closer, step by step, to a target. Google trained the model using many pairs of images, each including a person wearing a garment in two unique poses — for instance, an image of someone wearing a shirt standing sideways and another of them standing forward. To make the model more robust (i.e., combat visual defects like folds that look misshapen and unnatural), the process was repeated using random image pairs of garments and people. Google’s new AI-powered try-on feature, which taps generative AI to adapt clothing to different models. Starting today, U.S. shoppers using Google Shopping can virtually try on women’s tops from brands includi
- Google Maps gains glanceable directions and an expanded Immersive ViewGoogle Maps is introducing "glanceable directions" for walking, cycling and driving, allowing users to track progress from their lock screen, which is rolling out this month on Android and iOS. The feature also includes updated estimated times of arrival and notifications for turns. Google will also soon be saving places in Recent highlights, and will add tools to remove places that have been viewed, create custom routes and share places with others. Meanwhile, Immersive View is expanding to over 500 landmarks, with the ability to see weather and crowds at different times of day, along with peeks inside restaurants and cafes. 全文: TechCrunch Open Navigation TechCrunch Kyle Wiggers Timed to arrive ahead of the summer travel season, Google’s bringing new features to Maps focused on discovery and navigation. New “glanceable directions,” which Google announced in February, will start rolling out this month in the Maps app on Android and iOS. They’ll let users who opt-in track walking, cycling and driving trip progress from their lock screen or route overview, and see updated ETAs and notifications about where to make each turn. Image Credits: Google Arriving alongside glanceable directions is updates to Recents on desktop, a shortcut that automatically organizes places you researched in Maps into a highlight on the side of the Maps screen. Soon, Google will save places in Recent highlights even after you’ve closed the Google Maps window, and adds tools to remove places you’ve viewed, create custom routes between destinations and share a selection of places with someone else. Lastly, Immersive View, Google’s AI-powered feature on the iOS and Android Maps app that fuses together images to create a multidimensional view of the world, is expanding to over 500 landmarks. Among the highlights are Prague Castle and Sydney Harbour Bridge. Image Credits: Google Christina Tong, senior product manager at Google Maps, shared more in a blog post this morning: “With the ti
- Google intros new AI-powered travel and product search featuresGoogle has announced new features for its Search Generative Experience (SGE) search engine, with a focus on travel and shopping. When users ask a question about a place or destination, they will receive a snapshot that includes information from across the web, reviews, photos and business profile details. Shopping for a product will bring up details including product descriptions, reviews, ratings, prices, images and recommendations. The company said that its proprietary data, which includes details on over 200 million places and 35 billion product listings, sets it apart from rivals such as You.com and Perplexity AI. 全文: TechCrunch Open Navigation TechCrunch Kyle Wiggers Google’s new AI-powered search experience, released in May, might have gotten a mixed reception. But the search giant isn’t letting that slow its feature roadmap. Google today announced new capabilities, some of which it previously previewed at its I/O conference , heading to Search Generative Experience (SGE) — the moniker for its experimental search experience — focused on travel and shopping. Now, when a user asks questions about a place or destination in Google Search (e.g. “Is this restaurant good for large groups?”), they’ll see a snapshot that brings together information from across the web as well as reviews, photos and business profile details submitted by business owners. And when they’re shopping for a product — Bluetooth speakers, say — SGE will show factors to consider along with product descriptions, reviews, ratings, prices, images and recommendations. Searching a location using the new generative AI features in Google Search. The experiences evoke AI-powered search rivals like You.com and Perplexity AI . But Rany Ng, VP of Search, argues that they uniquely benefit from Google’s proprietary data, including details on more than 200 million places in the world and 35 billion product listings. A product search with SGE. “With this improved experience in SGE, you’ll get useful
June 14
- Netflix will open a restaurant that serves food from ‘Chef’s Table,’ ‘Nailed It!’ and moreNetflix is venturing into the food service industry by opening a pop-up restaurant called Netflix Bites. The restaurant will feature chefs from various Netflix cooking series, including “Chef’s Table,” “Nailed It!” and “Iron Chef: Quest for an Iron Legend”. Netflix Bites will only be open for a limited time, from June 30 to an unannounced date, and will be located in Los Angeles. The company has not revealed the menu or pricing. To make a reservation, a nonrefundable deposit of USD 25 per person is required and will go towards the final bill. 全文: TechCrunch Open Navigation TechCrunch Lauren Forristal Netflix is expanding beyond streaming and mobile gaming and into a completely new category–the food service industry. The company announced yesterday that it’s opening a pop-up restaurant called Netflix Bites, featuring chefs from various Netflix cooking series, such as “Chef’s Table,” “Nailed It!” and “Iron Chef: Quest for an Iron Legend.” Although the streaming giant has launched immersive experiences based on shows before, including “Stranger Things ,” “Squid Game ,” “Bridgerton” and “Money Heist,” this is the first time that Netflix is stepping into the restaurant business. Netflix Bites patrons will be able to taste food from “Iron Chef: Quest for an Iron Legend” chefs Curtis Stone, Dominique Crenn, Ming Tsai and Andrew Zimmern, as well as Rodney Scott and Ann Kim from “Chef’s Table,” and pastry chef Jacques Torres, the head judge from “Nailed It!” Other participating stars include Nadiya Hussain from “Nadiya Bakes” and mixologists from Netflix’s “Drink Masters”–Frankie Solarik, Julie Reiner, LP O’Brien and Kate Gerwin. Netflix Bites will only be open for a limited time. The restaurant will launch on June 30 and be open daily from 5-10 p.m. PT. There will also be a brunch menu available on the weekends from 10 a.m. to 2 p.m. Netflix didn’t reveal what exactly will be on the menu, nor did it disclose pricing. To book a reservation, people can visit www.net
- Europe takes another big step towards agreeing an AI rulebook欧洲议会投票确认了其AI法案的谈判授权,这是实现欧盟人工智能全面规则书的下一个阶段的重要里程碑。议员们支持了欧盟委员会提议的一个经过修订的版本,扩大了规则书,旨在确保在欧洲开发和使用的人工智能“完全符合欧盟的权利和价值观,包括人权监督、安全、隐私、透明度、非歧视和社会和环境福祉”。议员们已经支持的变化包括完全禁止远程生物识别监视和预测性警务,以及禁止“从互联网或闭路电视摄像头中无目标地抓取面部图像以创建面部识别数据库”——基本上是对Clearview AI及其类似公司的硬禁止。议员们还扩大了高风险AI系统的分类,包括那些对人们的健康、安全、基本权利或环境造成重大危害的AI系统,以及用于影响选民和选举结果的AI系统。议员们的提案还为消费者决策权增加了一系列消费者权利,包括如果AI系统给他们造成了伤害,消费者可以要求集体维权。 全文: TechCrunch Open Navigation TechCrunch Natasha Lomas The European Parliament has voted to confirm its negotiating mandate for the AI Act — hitting a major milestone which unlocks the next stage of negotiations toward a pan-EU rule-book for artificial intelligence. Parliamentarians backed an amended version of the Commission proposal that expands the rule-book in a way they say is aimed at ensuring AI that’s developed and used in Europe is “fully in line with EU rights and values including human oversight, safety, privacy, transparency, non-discrimination and social and environmental wellbeing”. Among the changes MEPs have backed is a total ban on remote biometric surveillance and on predictive policing. They have also added a ban on “untargeted scraping of facial images from the internet or CCTV footage to create facial recognition databases” — so basically a hard prohibition on Clearview AI and its ilk. The proposed ban on remote biometric surveillance would apply to both real-time or post (after the fact) applications of technologies like facial recognition, except, in the latter case, for law enforcement for the prosecution of serious crimes with judicial sigh off. MEPs also added a ban on the use of emotional recognition tech being used by law enforcement, border agencies, workplaces and educational institutions. Parliamentarians also expanded the classification of high risk AI systems to include those that pose significant harm to people’s health, safety, fundamental rights or the environment, as well as AI systems used to influence voters and the outcome of elections. Larger social medi
- Rejoice, for startup valuations are slowly recovering本文介绍了近期软件股票的涨势,以及公共软件公司的收入倍增。具体来说,投资者认为公共软件公司每个收入美元的价值更高,这将从某种程度上减轻初创公司筹集资金的压力。虽然恢复软件估值的努力还不足以消除初创公司流动性堆积的问题,但这是一个良好的开端。 全文: TechCrunch Open Navigation TechCrunch Alex Wilhelm We have good news for tech companies big and small this fine Wednesday morning: valuations are heading higher, though at a modest pace. New public market data indicates that software stocks are rising to their highest points so far this year. And to make it sweeter, the underlying revenue multiples at public companies are expanding as well, especially at the faster growing subset of software companies. The Exchange explores startups, markets and money. Read it every morning on TechCrunch+ or get The Exchange newsletter every Saturday. Investors today seem to think public software companies are worth more per dollar of revenue in aggregate. That will lift some pressure from startups struggling to raise new capital at attractive prices or harmonize legacy private-market companies’ valuations with their hoped-for IPOs. Clearing the massive startup liquidity backlog will take more than a modest reinflation of software valuations, but it’s a start. Starting with high-level numbers, here’s a look at the Bessemer Cloud Index’s performance this year. Note that the value of the constituent companies kicked off the year strong, gave back most of those gains, and then starting in May, began to pick up steam once again (YCharts data):
- Majority, a digital bank for US migrants, grabs $9.75M amid expansion in Texas美国移民数字银行Majority近日在开设社区聚会场所、扩大业务至得克萨斯州拉雷多市时完成了第四轮融资,融资金额为975万美元。该公司成立于2019年,以每月5.99美元的会员费向移民提供银行账户和借记卡、社区折扣、免费国际汇款和折扣国际通话等服务。该银行不需要社会保障号码或美国文件,只需要国际政府签发的身份证明和美国居住证明即可。其“顾问计划”为全国超过250名受过培训的支持人员网络,他们自己也是移民。该银行已经在佛罗里达州的迈阿密、海里亚和奥兰多以及休斯顿设有实体店。 全文: TechCrunch Open Navigation TechCrunch Christine Hall Majority , a mobile banking company offering financial services to U.S. migrants, closed on its fourth funding round in two years as it opens a community meetup space in Laredo, Texas. The new location positions the company close to The Gateway to the Americas International Bridge that crosses the Rio Grande, Magnus Larsson, founder and CEO at Majority, told TechCrunch. Majority already has physical locations across Florida, in Miami, Hialeah and Orlando, and in Houston. “We started off in Texas and Florida,” Larsson said. “Florida has been our main focus for a period of time now, however, our business is about trust, and there is an opportunity to be the first banking product that you can open when you enter the country. As an immigrant myself, I experienced how hard it can be to seek out the services and support you need.” For a $5.99 per month membership fee, the company, founded in 2019, offers migrants a bank account and debit card, community discounts, free international money transfer and discounted international calling. Accounts don’t require a Social Security number or U.S. documentation, just an international government-issued ID and proof of U.S. residence, and don’t have overdraft fees or minimum balance requirements. There is also access to Majority’s “Advisor Program,” a network of over 250 trained support staff nationwide, who are immigrants themselves. Majority, a mobile bank for migrants, closes third venture round in just over a year Larsson didn’t get specific on the company’s customer base or revenue beyond saying that Majority’s transaction volume is growing 6x per year. The $9.75 million in new capital is an extension to Majority’s $37.5 mil
- Backstrapped to 800,000 users, online programming community JDoodle lands fundingJDoodle, an Australian-based online programming community, has secured AUD3.2m ($2.2m) in funding led by Main Sequence, the venture firm founded by Australia’s national science agency. JDoodle currently has tools aimed at software developers of all levels, ranging from students to professionals, including zero-set up IDEs (integrated development environments) for more than 76 coding languages, with the goal of freeing up developers to focus on important tasks instead of small, unimportant ones. The new funding will go towards JDoodle’s plans to scale up to eight million users, including marketing, product development and hiring for its team in Australia. 全文: TechCrunch Open Navigation TechCrunch Catherine Shu Started as a side project, JDoodle , which lets developers practice more than six dozen programming languages on one website, hit 800,000 users while being bootstrapped. Now the Sydney, Australia-based startup has set a goal of 8 million users within the next 24 months after getting $3.2 million AUD (about $2.2 million USD) in new funding. The round was led by Main Sequence, the venture firm founded by Australia’s national science agency . JDoodle has tools aimed at software developers of all levels, ranging from students to professionals. These include zero-set up IDEs (integrated development environments) for more than 76 coding languages, including Java CloudIDE, that lets users write and run code in their browsers without having to set up local software or libraries. Over the past year, JDoodle hit 800,000 users, including 3,000 educational institutions. JDoodle was started by software developer Gokul Chandrasekaran, who spent seven years working on it and reached 500,000 users before quitting his job to focus on it full-time. It grew out of a side project with one page and programming language that Chandrasekaran set up for $20, before spending evenings and weekends adding new features based on user feedback. Chandrasekaran told TechCrunch that he
- Mitigating AI biases in healthcare with Amy Brown from Authenticx本文介绍了Authenticx的联合创始人兼首席执行官Amy Brown如何利用人工智能帮助保险公司和医疗组织从呼叫中心中提取数据,并减少人工智能在医疗保健领域中的偏见。她提到自己在呼叫中心工作的背景启发了她进军创业领域,同时也探讨了创业者的牺牲和非常人性化的一面,以及在构建AI模型时他们非常有意识的做法。 全文: TechCrunch Open Navigation TechCrunch Rebecca Szkutak Dominic-Madori Davis Welcome back to Found, where we get the stories behind the startups. This week Becca and Dom are joined by Amy Brown , the co-founder and CEO of Authenticx , a Midwestern startup that helps insurance companies and medical organizations extract data from their call centers using AI. Amy told Becca and Dom about how her background working in the same call centers inspired her to foray into entrepreneurship. She also talked about the sacrifices and very human side of being an entrepreneur and how they were very intentional when building their AI model. Subscribe to Found to hear more stories from founders each week. Connect with us: On Twitter On Instagram Via email: [email protected]
- Paro raises $25M to match freelance financial experts with firmsParo, a platform connecting freelance financial experts with businesses, has raised $25m in a series C funding round led by Top Tier Capital Partners. Paro uses algorithms to identify and recommend accountants and other financial experts to businesses, and customers can connect on-demand to pay them only for the work they need. The new funding will be used to develop Paro's platform, expand its customer base and support R&D. The company claims revenue has tripled since its series B round in Q3 2021 and freelancers are on track to bill over $1m through Paro this year. 全文: TechCrunch Open Navigation TechCrunch Kyle Wiggers In 2015, Michael Burdick and Dan Wywrot, who’d spent years working at large accounting firms, founded a startup — Paro — to deliver finance and accounting solutions to businesses through a network of freelance talent. They say they were inspired by the lack of flexibility in traditional accounting firms to choose which projects to work on — and when to work. Paro’s investors concur. Today, Paro announced that it raised $25 million in a Series C round led by Top Tier Capital Partners, bringing the startup’s total raised to $67 million. Madrona Venture Group, Revolution Ventures and Sierra Ventures also participated in what Paro CEO Anita Samojednik described as an up-round with “clean terms.” Samojednik told TechCrunch that the cash will be put toward further developing Paro’s platform, expanding its customer base and general R&D. “The overall goal of this round is to become the go-to resource for finance and accounting solutions and for businesses of all sizes, including enterprises,” Samojednik said. “Given the continued traction we have seen within our business, the ongoing embrace of remote work and the increasing need for flexible business solutions, now was the right time to raise and double down on our efforts to push our unique offering to more enterprises and the value we bring to businesses and experts.” Paro’s platform uses algor
- Informatica acquires Privitar, once valued at $400M, to expand its data management stack数据管理软件公司Informatica已收购英国初创公司Privitar,以扩展其数据管理堆栈。Privitar专注于构建和提供数据访问控制。此次收购将融入Informatica的更大数据堆栈中,后者将其作为“智能数据管理云”出售。Privitar成立于2014年,获得了超过1.5亿美元的融资,最后一笔融资是由华尔街投资公司Warburg Pincus领导的8000万美元C轮融资。Informatica的EVP和首席产品官Jitesh Ghai表示,他相信Informatica将进行更多的收购。 全文: TechCrunch Open Navigation TechCrunch Ingrid Lunden One of the recurring themes in the area of enterprise IT these days is consolidation, and today we’re getting another dose of it in the area of enterprise data management. Informatica — which provides a set of tools to enterprises to analyze, manage and share data across their organizations — is acquiring Privitar, a startup out of the U.K. that focuses on building and providing data access controls. The acquisition will sit within Informatica’s larger data stack, which it sells as the “intelligent data management cloud.” Terms of the deal are not being disclosed. Privitar notably had raised over $150 million in funding since first being founded back in 2014, from investors that included Accel and Warburg Pincus as well as strategics like Salesforce, HSBC and Citigroup. Its last valuation was over $400 million. But that was all the way back in 2020, at the peak of the last tech fundraising and valuation cycle, when Privitar had last raised outside funding, an $80 million Series C led by Warburg Pincus. Valuations have seen a lot of downward pressure in the current market, and companies have found it a lot more challenging to close rounds even at less attractive terms. Add that to another surprise — we are not seeing the expected surge of M&A that many thought there would be this year (to make up for the lack of funding) — the story right now seems to be: a lot less competition for deals, and a buyer’s market. Consolidation has been playing out across the board in enterprise IT. At one end, customers are looking to reduce the number of suppliers bring down costs and make things more efficient; at the other, smaller tech companies, running out of funding and
- OctoML launches OctoAI, a self-optimizing compute service for AIOctoML推出OctoAI,一种自我优化的AI计算服务,旨在帮助企业使用现有的开源模型,并使用自己的数据对其进行微调,或使用该服务托管自己的自定义模型。OctoAI是一种自我优化的计算服务,可帮助企业构建基于机器学习的应用程序,并将它们投入生产,而无需担心底层基础设施。用户只需决定他们想优先考虑什么(延迟还是成本),OctoAI将自动为他们选择合适的硬件。该服务还将自动优化这些模型(从而实现额外的成本节省和性能提升),并决定是否在Nvidia GPU或AWS的Inferentia机器上运行它们。这大大简化了将模型投入生产的复杂性,这仍然是许多机器学习项目的障碍。OctoML还提供了流行基础模型的加速版本,如Dolly 2、Whisper、FILM、FLAN-UL2和Stable Diffusion等。值得注意的是,虽然OctoML将继续与现有客户合作,这些客户只想使用该服务来优化其模型,但该公司未来的重点将放在这个新的计算平台上。 全文: TechCrunch Open Navigation TechCrunch Frederic Lardinois When OctoML launched in 2019, its primary focus was optimizing machine learning (ML) models. Since then, the company has added features that make it easier to deploy ML models (and raised $132 million ). Today, the company is launching the latest iteration of its service — and while it’s not quite a pivot, it does shift the company’s emphasis from optimizing models to helping businesses use existing open-source models and fine-tune them with their own data or use the service to host their own custom models. The new OctoML platform — dubbed OctoAI — is a self-optimizing compute service for AI, with a special emphasis on generative AI, that helps businesses build ML-based applications and put them into production without having to worry about the underlying infrastructure. “The previous platform was focused on ML engineers and optimizing and packaging the models into containers that could be deployed across different sets of hardware,” OctoML co-founder and CEO Luis Ceze explained. “We learned a ton from that, but the next natural evolution is to have a fully managed compute service that abstracts all of that [ML infrastructure] away.” Image Credits: OctoML With OctoAI, users simply decide what they want to prioritize (think latency vs. cost) and OctoAI will automatically choose the right hardware for them. The service will also automatically optimize these models (leading to additional cost savings and performance gains) and decide whether it’s best to run them on Nvidia GPUs or AWS’s