<ul><li>Federico Viticci has developed S-GPT (Shortcuts-GPT), a shortcut that integrates <keyword id="15681942" type="General" weightage="20" keywordseo="chatgpt" source="Orion">ChatGPT</keyword> into iOS, <keyword id="7719460" type="General" weightage="20" keywordseo="ipados" source="Orion">iPadOS</keyword>, macOS, and watchOS for Apple users.</li><li>S-GPT leverages OpenAI's API to unleash the power of ChatGPT, making it more accessible and user-friendly for tech enthusiasts and developers alike.</li><li>Below, read about simple steps to use S-GPT.</li></ul>Software developer and a renowned technology enthusiast, Federico Viticci, has developed a game-changing shortcut, S-GPT (Shortcuts-GPT), which allows users to integrate ChatGPT into the native system features of iOS, iPadOS, macOS, and watchOS. While the shortcut is free for all, users will need to set up an OpenAI pay-as-you-go billing plan to access its full potential. Leveraging Open AI's developer API, S-GPT unleashes the power of ChatGPT's latest API, though it can also be used with the existing ChatGPT 3.5 model or ChatGPT 4 for those with access. With S-GPT, Viticci aims to simplify the process of integrating ChatGPT into various Apple platforms, making it more accessible and user-friendly for tech enthusiasts and developers alike. Get S-GPT from the <a href="https://www.macstories.net/ios/introducing-s-gpt-a-shortcut-to-connect-openais-chatgpt-with-native-features-of-apples-operating-systems/">MacStories</a> website. Set it up on your iPhone, iPad, or Mac using the Shortcuts app. I highly recommend it giving it a try. S-GPT utilizes the ChatGPT API, which requires payment for usage, but the costs are minimal at $0.002 per 1,000 tokens (tokens are pieces of words used by the chatbot for processing prompts). According to Viticci, he has only spent $1.50 (₹123) so far while using S-GPT. <cmsannotation type="cmsannotation" annotationid="How to set up S-GPT" comment="subhead" text="How to set up S-GPT">How to set up S-GPT</cmsannotation> <ol><li>Create an OpenAI account and make sure you have a pay-as-you-go billing plan.</li><li>Get a ChatGPT API Key from your OpenAI account settings by selecting ‘Create a new secret key.’</li><li><a href="https://www.icloud.com/shortcuts/882c9a2870c0431098905445a5f1511e">Download</a> the S-GPT shortcut onto your device and start the setup process.</li><li>When prompted for an API key, paste in the one you copied earlier.</li><li><a href="https://www.icloud.com/shortcuts/b66b86615b0b478bad95948616e4f48a">Download</a> the S-GPT Encoder.</li><li>If everything is set up correctly, the S-GPT shortcut should work as expected, even with Siri.</li></ol>S-GPT has many applications, including summarizing webpages in Safari through Share Sheet, analyzing text in your clipboard using the trigger word ‘clipboard,’ keeping track of your reminders and calendar with the trigger word ‘my schedule,’ and creating customized playlists in the Music app with the trigger word ‘playlist.’ SEE ALSO: <a href="https://www.businessinsider.in/tech/news/bard-to-magi-googles-latest-weapon-in-the-battle-for-search-supremacy/articleshow/99576251.cms">From Bard to Magi: Google's latest weapon in the battle for search supremacy</a> <a href="https://www.businessinsider.in/tech/news/ai-generated-images-and-videos-a-game-changer-or-a-threat-to-authenticity/articleshow/99560443.cms">AI-generated images and videos: A game-changer or a threat to authenticity?</a>

You can now get ChatGPT on your iPhone through this shortcut - here’s how

<ul class="summary-list"><li>Universities across the globe are scrambling to draft AI policies before the new academic term.</li><li>But some educators say the real threat isn't AI, but a "lagging and outdated approach to education."</li></ul>As the new school year approaches, universities are scrambling to draft new policies for a post-ChatGPT classroom.<a href="https://www.businessinsider.com/chatgpt-schools-colleges-ban-plagiarism-misinformation-education-2023-1">Some schools have banned the tool entirely</a>, over fears it would enable AI-assisted plagiarism. Some professors have <a href="https://www.businessinsider.com/chatgpt-driving-return-to-paper-exams-written-essays-at-universities-2023-8">opted to return to paper exams to fight students using ChatGPT</a>, while one university even resorted to springing <a href="https://www.hindustantimes.com/cities/bengaluru-news/why-this-bengaluru-institute-has-restricted-chatgpt-use-for-students-101674901758231.html">surprise checks on students suspected of using the tool</a>.Madison White, a student at Stetson University in Florida, <a href="https://www.bloomberg.com/news/articles/2023-08-15/how-will-colleges-handle-chatgpt-in-new-school-year-it-s-a-work-in-progress" target="_blank" rel="noopener">told The Wall Street Journal about professors' widespread fears of AI-assisted cheating</a>. "They often immediately assumed that it was a hack for students to get away from doing readings or homework," she said.Insider spoke to six educators at universities based in Singapore, the Netherlands, Hong Kong, and the United States. Four of them said that fears over the use of AI in classrooms were overblown, and all six said schools can't afford to shut ChatGPT out of classrooms entirely.One of them, Ian Chong, an associate professor of international relations at the National University of Singapore, said that he wasn't losing sleep about AI seeping into education."I like to tell my students if they use AI to write, I'll use AI to give comments and write their recommendation letters," he said. <h2>Schools shouldn't overly on AI detection tools, the educators said</h2>The concern over AI-assisted cheating has fueled a demand for tools to detect AI-generated content, but the would-be solution has introduced a new set of problems.Rebecca Tan, a political science lecturer at the National University of Singapore, told Insider AI detection tools can be "notoriously inaccurate."As of July, plagiarism-checking software Turnitin's AI detection tool <a href="https://www.businessinsider.com/openai-chatgpt-ai-detection-tool-shut-down-over-inaccuracy-2023-7">had been used to review over 65 million papers</a> since its launch in April — <a href="https://www.washingtonpost.com/technology/2023/04/01/chatgpt-cheating-detection-turnitin/">despite concerns over its accuracy</a>. Also in July, ChatGPT creator OpenAI <a href="https://www.businessinsider.com/openai-chatgpt-ai-detection-tool-shut-down-over-inaccuracy-2023-7">quietly shut down its AI detection tool</a> over concerns about its inaccuracy.Instead of relying on AI detection tools, educators need to get innovative as AI tools become ubiquitous — through ideas like having students submit the introductions to their essays first, Tan said."It's short so students don't feel like they need to use ChatGPT, but it gives me a point of reference when I'm marking their longer essay to see if the final product deviates greatly from the initial introduction," she said of the tactic.Educators are also rethinking how students are assessed. Chong said he's using more "complex, scenario-based" exercises with students, for which ChatGPT's responses remain inadequate. Michael Rivera, a lecturer at Hong Kong University's history department, recommends using live discussions and reflections as assessments.But, "pursuing such an approach is only realistic if class sizes are kept small enough to pay personal attention to everyone and provide detailed feedback throughout the school term," said Shannon Ang, an assistant professor of sociology at Singapore's Nanyang Technological University, who was critical of the pressures faced by universities to expand class sizes in the name of "efficiency."Plus, an emphasis on getting around AI might be missing the point.<h2>Schools can't afford to ignore teaching with AI</h2>Educators Insider spoke to agreed that universities shouldn't shut ChatGPT out of classrooms entirely, as ignoring the chatbot might do more harm than good."What I do, for example, is to let students use AI tools to answer certain research questions, and then compare it to Wikipedia — their sort of second love in the pre-AI days — before comparing it to their textbooks," said Kai Jonas, a professor of applied social psychology at Maastricht University in the Netherlands.Jonas said the key is to embrace AI tools in the classroom and teach students how to fact-check ChatGPT's responses."We've already seen cases of AI generating false sources of information, and this would be a useful example to interrogate," said Joana Cook, an assistant professor of political violence at Netherlands' Leiden University, currently lecturing at Johns Hopkins University.The need to vet information from AI stems from the technology's ability to <a href="https://www.businessinsider.com/how-ai-chatbots-like-chatgpt-work-explainer-2023-7">"hallucinate," or present wrong information as fact</a>. High-profile instances of this include the technology making up fake court cases to reference — <a href="https://www.businessinsider.com/chatgpt-generative-ai-law-firm-fined-fake-cases-citations-legal-2023-6">leading to a law firm receiving a $5000 fine</a> — and <a href="https://www.businessinsider.com/microsoft-ai-written-article-recommend-food-bank-tourist-attraction-2023-8">retractions by media outlets over errors made by AI</a>.<h2>The real threat to education isn't AI, it's boring lessons</h2>When asked about threats to education, Nanyang Technological University's Ang said: "AI tools are not the threat — a lagging and outdated approach to education is."Educators who take the time to provide thoughtful feedback are likelier to see engaged students who do the work, as students understand that "it is pointless to receive feedback on work that is not your own," said Ang.That's because using AI tools like ChatGPT as a shortcut hints at students being unengaged or swamped, and thus not seeing the value of doing the work, echoed the National University of Singapore's Tan.It takes work from professors and educators to overcome inertia and adapt to the technology while meeting students' needs, Jonas said: "I can see that some of them are hesitant to do so because they have to revise their curricula and teaching materials.""I believe it is 100% the success or the failure of the teacher if students don't know how to use AI tools strategically in their education," said Hong Kong University's Rivera.

The GPT college classroom is almost here, but professors say AI isn't the real threat

<ul class="summary-list"><li>Workers using OpenAI's ChatGPT may actually perform more poorly than those who don't, new BCG research finds.</li><li>That's because some workers take the AI chatbot's outputs at face value and don't check for errors.</li></ul>If you're using <a target="_blank" href="https://www.businessinsider.com/ai-tools-make-your-job-easier-boost-productivity-work-chatgpt-2023-5">ChatGPT as a shortcut at work</a>, you may want to be extra careful.Using AI at work can actually hurt your job performance — if it's used for tasks outside of its capabilities, <a target="_blank" href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4573321&amp;utm_source=newsletter&amp;utm_medium=email&amp;utm_campaign=newsletter_axiosam&amp;stream=top">new research from Boston Consulting Group found</a>.A group of researchers from BCG, Harvard, Wharton, and MIT conducted an experiment to see how access to AI impacts white-collar workers' productivity and quality of work.To test this, researchers randomly assigned 758 consultants at BCG across one of three groups: one with no access to AI; one with access to <a target="_blank" href="https://www.businessinsider.com/gpt-4-openai-model-api-details-2023-3">ChatGPT powered by GPT-4</a>; and one with access to ChatGPT, as well as instructional videos and documents on prompt engineering strategies.After establishing performance baselines, consultants in each group were assigned one of two categories of tasks.One category included 18 tasks that exist "inside the frontier" of what the AI can do, like brainstorming innovative beverage concepts or coming up with a thorough business plan for a new footwear concept.The other category contained more open-ended tasks that exist "outside the frontier" of AI's capabilities. While "consultants would excel" at these tasks, "AI would struggle without extensive guidance," the study said. For example, consultants assigned this set of tasks were asked to offer recommendations to the CEO of a hypothetical company by using internal financial data and interviews with company insiders — information the AI didn't have access to.Researchers found stark differences in the results of the three groups, depending on their access to ChatGPT.For tasks "inside the frontier," consultants using AI were "significantly more productive" and "produced significantly higher quality results" than those who weren't using the chatbot. However, consultants using AI to complete tasks "outside the frontier" were "19 percentage points less likely to produce correct solutions compared to those without AI." That's because the consultants with AI were found to indiscriminately listen to its output — even if the answers were wrong.These findings demonstrate AI's "uneven" capabilities.While the study's findings show that AI is "exceedingly good" at helping humans with some tasks, humans should exercise caution when using the technology to avoid errors, Saren Rajendran, one of the researchers involved in the study, told Insider in an email."We should be mindful when using GenAI," he added.BCG's findings demonstrate a cautionary tale for workers thinking about using ChatGPT to help do their jobs. Since ChatGPT came out last November, workers across industries have been <a target="_blank" href="https://www.businessinsider.com/ai-chatgpt-jobs-side-hustle-make-more-money-accessible-2023-8">using the AI chatbot</a> — <a target="_blank" href="https://www.businessinsider.com/chatgpt-secret-productivity-work-ai-technology-ban-employees-coworkers-job-2023-8">sometimes without telling their bosses</a> — to <a target="_blank" href="https://www.businessinsider.com/chatgpt-amazon-employees-use-ai-chatbot-software-coding-customer-questions-2023-1">develop code</a>, <a target="_blank" href="https://www.businessinsider.com/how-use-ai-beginner-easy-increase-productivity-2023-8">create marketing materials</a>, and <a target="_blank" href="https://www.businessinsider.com/high-school-math-science-teacher-uses-openai-chatgpt-education-learning-2023-3">generate lesson plans</a>.However, ChatGPT's outputs aren't perfect and can contain "<a target="_blank" href="https://www.businessinsider.com/ai-chatbot-chatgpt-google-microsofty-lying-search-belief-2023-2">hallucinations</a>."Tech publication CNET was put on blast earlier this year after readers noticed that a number of its <a target="_blank" href="https://www.businessinsider.com/tech-site-issued-corrections-after-ai-writing-got-facts-wrong-2023-1">AI-generated articles included factual errors</a>.As of September 28, media watchdog <a target="_blank" href="https://www.newsguardtech.com/special-reports/ai-tracking-center/">NewsGuard</a> has identified 487 "unreliable" AI-generated news sites with "little to no human oversight." In an ad for Google's Bard, <a target="_blank" href="https://www.businessinsider.com/google-ad-ai-chatgpt-rival-bard-gives-inaccurate-answer-2023-2">the AI chatbot made a factual error</a> when asked about the James Webb Space Telescope.AI-generated errors may only get worse: In a recent paper, AI researchers found that <a target="_blank" href="https://www.businessinsider.com/ai-model-collapse-threatens-to-break-internet-2023-8">generative AI models could soon be trained on AI-generated content </a>— a phenomenon they call "model collapse." The result could be more low-quality outputs in the near future."As the boundaries of AI capabilities continue to expand, often exponentially, it becomes incumbent upon human professionals to recalibrate their understanding of the frontier and for organizations to prepare for a new world of work combining humans and AI," the researchers wrote.

OpenAI's ChatGPT can actually make workers perform worse, a new study found

<ul class="summary-list"><li>Keith Peiris is the co-founder of <a href="https://beta.tome.app/?utm_source=google&amp;utm_medium=cpc&amp;utm_campaign=Feb2023_T1_Google-SEM_Prospecting_Tome-Brand&amp;utm_content=Feb2023_T1_Google-SEM_Prospecting_Tome%20AI&amp;utm_ad=649586285008&amp;utm_term=tome.ai&amp;matchtype=b&amp;device=c&amp;GeoLoc=9004057&amp;placement=&amp;network=g&amp;campaign_id=19009600242&amp;adset_id=143821231677&amp;ad_id=649586285008&amp;gclid=CjwKCAjw8-OhBhB5EiwADyoY1QetLEkFF3nLyW4nKL8L-OuAouDk9ijTPUtI63giup3r-xg7ERRuOBoCJ3MQAvD_BwE" target="_blank" rel="noopener">Tome</a>, a storytelling tool that uses generative AI models like GPT-4 and Dall-E-2.</li><li>The platform hit 1 million users less than 5 months after launching, eclipsing tools like Dropbox and Slack.</li></ul>This as-told-to essay is based on a conversation with Keith Peiris, cofounder of Tome. It has been edited for length and clarity.You could say I've been in the business of communication and storytelling my whole life. When I was nine years old, my dad gave me a giant book on Photoshop. He worked at an IT company selling computers, and he eventually got a hold of a bootlegged copy of Photoshop for me to experiment with. This was the early 2000s, so if you wanted to build animations on the web, you were using Macromedia Flash. I started playing around with all these tools and posting my creations on forums. Soon, people started messaging me. Then they started commissioning me to build websites for them. I was charging $250 for a website, and I quickly realized that with all the money flowing into my account via PayPal, I would have no problem affording an Xbox. I had a bit of a flywheel going by the time my dad caught wind of my web-building venture. He suggested that we turn the operation into a full-fledged company, and that I take the helm as CEO. We named our company Cyberteks Design. Eventually, my dad quit his job and jumped on board with me full-time. At our peak, we had around 100 clients and were charging $5,000 a website. We weren't valued like a tech company, of course. At the age of 15, though, I had become the accidental breadwinner of our household. That gave me the confidence that I could make it as an entrepreneur in the long run. <h3>'You're just building products for people'</h3>I went off to college first, studying nanotechnology engineering at the University of Waterloo. My first job out of college was at Facebook. I spent more than six years there, including a couple of years at Instagram and a brief stint at Oculus, before becoming an entrepreneur-in-residence at the venture firm Greylock. My biggest takeaway from those years in big tech was that even at a company with a global reach like Facebook you're still just building products for people. At the same time, I also learned that no one uses your tools the way you envisioned. When I was working on direct messages, now known as "DMs" at Instagram, my team and I thought we were building a tool for people to share memes with close friends. Instead, we noticed that people were using the feature for everything from selling their personal apparel to tracking celebrities to striking up new friendships with people who shared their interests. The truth is that everyone uses tools differently, so you need to build them in a permissive way. <h3>The same has been true with Tome</h3><div type="image" title="Reid Hoffman" height="1281" width="2562" msid="99516440" hostid="261" sizebytes="41938"></div><a href="https://beta.tome.app/?utm_source=google&amp;utm_medium=cpc&amp;utm_campaign=Feb2023_T1_Google-SEM_Prospecting_Tome-Brand&amp;utm_content=Feb2023_T1_Google-SEM_Prospecting_Tome%20AI&amp;utm_ad=649586285008&amp;utm_term=tome.ai&amp;matchtype=b&amp;device=c&amp;GeoLoc=9004057&amp;placement=&amp;network=g&amp;campaign_id=19009600242&amp;adset_id=143821231677&amp;ad_id=649586285008&amp;gclid=CjwKCAjw8-OhBhB5EiwADyoY1QetLEkFF3nLyW4nKL8L-OuAouDk9ijTPUtI63giup3r-xg7ERRuOBoCJ3MQAvD_BwE" target="_blank" rel="noopener">Tome</a> is a multimodal storytelling tool that utilizes generative AI. It's the first of its kind to allows creators to generate narratives ranging from presentations to outlines to stories along with visuals within seconds.Type a command like "Make a fundraising pitch for a company that makes telescopes," and Tome will roll out eight slides with a table of contents and pictures to match. When we started building Tome in the early days of the pandemic, my cofounder, Henri Liriani, and I thought we were building for someone we knew. Our archetype was a technical software designer, or a product manager, who had ideas to present. But since we launched in September, we've seen Tome being used by parents to create stories for their children. Doctors are making pre-surgery explanations and post-operation guides. Greylock partner and LinkedIn founder Reid Hoffman made a visual podcast with Tome. One user created a guide for smoking a Texas brisket. My mom used to have a WhatsApp group with her Sri Lankan "auntie" friends that sort of functioned like a giant photo gallery. Now, they're using Tome, which is a use case I had never even thought about. In February, we became the fastest productivity tool to hit 1 million users, outpacing platforms like <a href="https://www.businessinsider.com/slack-reaches-1-million-concurrent-users-2015-10">Slack</a> and <a href="https://www.businessinsider.com/one-million-users-startups-2012-1">Dropbox</a>, according to news reports. As of March, we've accrued 3 million users. After our Series B round led by Lightspeed Venture Partners, we were valued at $300 million. We didn't take any shortcuts to get here. And honestly, I don't know if there are any shortcuts to building a valuable product that people love. I think you have to be committed to doing things the hard way — the earnest way. If you do, though, I tend to believe that things work out.

I've been building popular tech products like Tome since I became my family's breadwinner at 15 years old. I've learned you never know how people are going to use your creation.

<ul><li>Needl - a cross-platform application that lets you search across your local file system and accounts.</li><li>Needl can be integrated with <keyword id="51500" type="General" weightage="100" keywordseo="gmail" source="Orion">Gmail</keyword>, <keyword id="1815769" type="General" weightage="20" keywordseo="slack" source="Orion">Slack</keyword>, <keyword id="6412891" type="General" weightage="20" keywordseo="google-drive" source="Orion">Google Drive</keyword>, and more.</li><li>It is available for both Windows and <keyword id="62949" type="General" weightage="50" keywordseo="mac" source="Orion">Mac</keyword>.</li></ul>Have you ever struggled to find a document and couldn’t remember if it was shared over Gmail, Slack, or Skype? Needl, lets you search across your multiple accounts and helps you find a required document within seconds. Google, MSN, Bing, DuckDuckGo, and other search engines help you to find information from the web, but it is hard to find documents, emails, messages, and meeting schedules from your own accounts. <a href="https://www.needl.tech/">Needl</a> is a cross-platform application that lets you search across your local file system and accounts like Google Drive, Gmail, Google Calendar, Slack, Skype, and more. The free version of Needl is available on the web, Windows, and Mac websites and lets you connect a single account per integration. In the free version, you will get limited access and standard integration to Slack, Gmail, Notion, Google Drive, and Calendar. If you need more features and integrations like Jira and Linear, then you have to pay $10 per month, which you get multiple accounts per platform and 24/7 customer support. <cmsannotation type="cmsannotation" annotationid="How to use it?" comment="subhead" text="How to use it?">How to use it?</cmsannotation> Needl is easy to set up and use, once installed, it will ask you to connect your accounts like Google, Slack, and more. Once done, you can easily search for files, events, emails, and other things across all these accounts and your local file system. Also, you can filter your search by messages, files, events, emails, and tasks. <div type="image" title="Needl" msid="95119043" hostid="261" sizebytes="35396"></div> The app also comes in handy with different shortcuts for navigating the app. Further, you can also customize the view to get more contextual information. Since it’s a matter of data and access to private accounts, the company clarifies all the information is stored on an AWS server and uses SSL to encrypt data during transit and 256-bit AES encryption at rest. Further, Needl claims that ‘We will never sell your data.’ Max Keenan, Angela Liu, and James Liu are all Needl founders and alums of the <keyword id="142498" type="General" weightage="50" keywordseo="university-of-chicago" source="Orion">University of Chicago</keyword>. They met at a hackathon and worked on projects like a tool to write essays using GPT-2 and TikTok for blog posts. The company has raised $2.5 million from different investors. Needl was rolled out to 200 users under a closed beta in August, and now the company is making it available to everyone. SEE ALSO: <a href="https://www.businessinsider.in/tech/news/this-website-generates-videos-from-a-single-image-and-text-for-free/articleshow/94561711.cms">This website generates videos from a single image and text for free – here’s all you need to know</a> <a href="https://www.businessinsider.in/science/news/soon-you-might-be-able-to-touch-and-feel-the-online-products-through-your-smartphone/articleshow/94961160.cms">Soon you might be able to touch and feel the online products through your smartphone</a>

Search engine for your online accounts — Here's how to find anything from your Gmail, Slack, Messenger and similar at once

<ul class="summary-list"><li>The Cloud 1.0 era was dominated by Amazon Web Services. </li><li>The generative AI boom has ushered in the Cloud 2.0 era. </li></ul>When you're building a house, you can do it in various ways. You can build all the tools you need and then hike into a forest to chop down wood and turn that into your final structure. Or, you can go to Home Depot and buy wood that's already been chopped down, along with all the nails, hammers and other equipment you need.Or, you can pay a contractor to do most of the work. Over the past 15 years, <a target="_blank" href="https://www.businessinsider.com/amazon-aws-struggles-in-saas-business-applications-cloud-market-hubspot-2021-11">Amazon Web Services</a> became the go-to contractor for anyone wanting to build something online. A website, an app, a business software service. Nowadays, very few developers venture into the digital forest and chop down their own semiconductors and build their own servers and data centers. AWS does it all for you, and provides most of the tools you need. This is what turned AWS from a fringe idea in 2006 into a $80 billion business with juicy profit margins and a huge, sticky customer base. This is the essence of a technology platform. The rise of <a target="_blank" href="https://www.businessinsider.com/generative-ai-threat-to-saas-companies-software-as-a-service-2023-5">generative AI</a> has created what some technologists are calling Cloud 2.0. There's a new stack of hardware, software, tools, and services that will power AI applications for years to come. Which company will be the go-to contractor in this new world? When you're building your AI house, who will you turn to for all that digital wood and the right tools that just work well together to get the job done?At the moment, AWS doesn't look like the leader here, while <a target="_blank" href="https://www.businessinsider.com/nvidia-shocker-1-trillion-spent-ai-data-centers-4-years-2023-8">Nvidia</a> has a massive head start. <div type="image" title="jensen huang drive nvidia ces" height="1274" width="2000" msid="104157223" hostid="261" sizebytes="151410"></div>Nvidia designs and sells <a target="_blank" href="https://www.businessinsider.com/tech-giants-spending-nvidia-gpus-google-meta-amazon-2023-7">GPUs</a> that are the go-to chip used by almost everyone who is creating large language models and other big AI models. These components were being used by AI researchers more than a decade ago when <a target="_blank" href="https://www.businessinsider.com/andrew-ng-ai-artificial-intelligence-playbook-2018-12">Andrew Ng</a> and others discovered GPUs could handle many computations in parallel. In contrast, AWS launched its Trainium GPU for AI model training in December 2020, almost a decade later. During that decade, Nvidia also built a platform around its GPUs that helped AI developers do their jobs. The company created compilers, drivers, an operating system, tools, code libraries, training notes and videos, and more. The core of this platform is a programming language called CUDA. There are now more than 4 million registered Nvidia developers who use CUDA all its related tools. That's up from 2 million in August 2020, and 1 million two years before that. The closest thing AWS has right now is called Neuron. This was released in 2019, according to <a target="_blank" href="https://github.com/aws-neuron/aws-neuron-sdk/releases/tag/v1.0.0">its Github page</a>. CUDA came out in 2007. <h2>Libraries, models, and frameworks</h2>Michael Douglas, a former Nvidia executive, spoke to Bernstein analysts recently about all this, explaining that the chipmaker's software and tools are far ahead of the competition: <ul><li>There are 250 software libraries available to developers that sit on top of CUDA. A library is pre-built code that performs a function like handling video or audio or doing something in graphics.</li><li>There are also 400 pre-built models that Nvidia offers developers for AI tasks such as image recognition and voice synthesis.</li><li>These libraries and models are weaved together with popular software frameworks such as PyTorch, TensorFlow, Windows and VMware. </li><li>This all sits in stacks on top of 65 platforms. A platform provides most of the other software developers need to operationalize something in an environment in a data center, or at the edge, with AI. </li></ul>"What that means is that, it's like if you want to build a house, the difference between going into the forest and having to cut down your own trees and make your own tools or going into Home Depot, where you buy the lumber and the tools and get everything to work together," Douglas said. <h2>Cloud 2.0</h2>Another key point here: Most AI developers already know how to use CUDA and Nvidia GPUs. All the hacks, tricks, and other shortcuts are embedded in the community. Arguably, Nvidia has already created an AI cloud platform – as AWS once did for the Cloud 1.0 era. As we move into Cloud 2.0, this will be hard for AWS, or anyone else, to break. "Nvidia has been betting on AI for a long time. They have a level of software maturity that no other provider can match," said Luis Ceze, CEO and cofounder of AI startup OctoML. "With Amazon, there's still a lot to mature there. People must have confidence that the silicon is good and the system software is there." Amazon's offering is just not as complete yet, he added. "Look at the experience of developers trying to bring AI models to train on AWS versus Nvidia. These are world's apart right now," Ceze explained.<h2>The Amazon-Anthropic feedback loop</h2>Don't count AWS out yet, though. It's early days, as <a target="_blank" href="https://www.businessinsider.com/andy-jassy-amazon-web-services-ceo-2021-1">Andy Jassy</a> and <a target="_blank" href="https://www.businessinsider.com/aws-ceo-adam-selipsky-stressed-about-amazon-day2-culture-2023-7">Adam Selipsky</a> keep saying. <a target="_blank" href="https://www.businessinsider.com/amazon-vp-distinguished-engineer-james-hamilton-profile-aws-jassy-boat-2021-8">James Hamilton</a> is an AWS cloud infrastructure genius who can take on Nvidia, even if the chipmaker has a major head start. "Nvidia has a lock on the market due to advantages in software," said Oren Etzioni, a technical director at the AI2 Incubator and a partner at Madrona Venture Group. "There's an argument that this window is starting to close."The Cloud 2.0 race really kicked off in late September when Amazon announced an <a target="_blank" href="https://www.businessinsider.com/amazon-invests-4-billion-in-anthropic-2023-9">investment of up to $4 billion</a> in an AI startup called <a target="_blank" href="https://www.businessinsider.com/anthropic-cofounder-daniela-amodei-trust-safety-generative-artificial-intelligence-2023-4">Anthropic</a>. This relatively small San Francisco-based company is packed with researchers who helped create the core GPT technology at the heart of today's generative AI boom. Anthropic has one of the most capable large AI models out there, called Claude.<div type="image" title="Dario Amodei Anthropic" height="5179" width="5179" msid="104157225" hostid="261" sizebytes="2214932"></div>Anthropic will now use AWS's in-house Trainium and Inferentia chips to build, train, and deploy its future AI models, and the two companies will collaborate on the development of future Trainium and Inferentia technology.A key part here is the feedback loop between Anthropic and AWS. Those AI pioneers at Anthropic are now committed to using AWS's GPUs, along with all the related software and tooling. If AWS is lacking here, Anthropic can help identify the biggest holes and suggest ways of filling them. To catch Nvidia, AWS really needed a marquee developer like this. All it took was $1 billion to $4 billion. "Amazon needs a major AI player using its own silicon," said Ceze. "Amazon being able to show Anthropic running their workloads on Amazon chips is a win. They want other folks building AI models to feel confident that the silicon is good and the software is mature."<h2>Avoiding the Nvidia tax</h2>Making your own AI chips is a really important advantage for winning the Cloud 2.0 race. If AWS has to keep buying Nvidia GPUs, it will have to pay a margin to Nvidia. (Some AI experts call this the Nvidia tax). That will make AWS's Trainium and Inferentia cloud services more expensive to run. If, instead, AWS can make its own AI chips, then it gets to keep that Nvidia profit margin for itself. That means it can offer AI training and inference cloud services that are more cost effective. At a recent AWS event in Seattle, Selipsky, AWS's CEO, hammered this message home: AWS's AI cloud services will be competitive based on "price performance." This means it should be more efficient to build AI models on AWS, according to the CEO. <div type="image" title="Adam Selipsky reInvent" height="1996" width="3000" msid="104157227" hostid="261" sizebytes="115550"></div>In the Cloud 1.0 era, AWS pulled this off partly through the acquisition Annapurna Labs in 2015. This startup helped Amazon design its own Gravitron CPUs for its cloud data centers. Suddenly, AWS no longer needed to buy as many expensive Xeon chips from Intel. This was an important element that made AWS such a competitive and successful cloud provider. <h2>How easy it is to switch from Nvidia?</h2>Here's the problem though: You can have great AI chips and great prices, but will anyone switch from using Nvidia GPUs and the CUDA platform? If they don't, AWS might not win Cloud 2.0 in the way it did with Cloud 1.0. This brings us back to the house-building analogy. At the moment, Nvidia probably offers the best combination of wood and tools to get the job done. If you're already using this, can you switch easily? I asked Sharon Zhou, cofounder of Lamini AI. Her startup spent months building a data center from scratch to help customers train AI models. When it was unveiled in September, the surprise was that Lamini used AMD GPUs rather than Nvidia chips. How was the switch? Zhou said Lamini didn't even try because it would have been so hard. "We never switched. We started the company on AMD. That's why it works," she said. "Switching would be more difficult than starting from scratch. If you want to switch, you'd have to apply a major refactor to your entire codebase."She also went back to one of the same points that Douglas, the former Nvidia exec, made: The AI libraries and other frameworks and tooling around CUDA are just more developed. "Libraries are heavily integrated with Nvidia today," Zhou said. "You need to reimplement fundamental abilities to work with large language models. Outside of the fundamentals, you need to do so for other common libraries, too."<h2>Nvidia's lock-ins</h2>Etzioni said Nvidia still has 2 main lock-ins with AI developers. The first is that PyTorch, a popular AI framework, is optimized really well for CUDA and therefore Nvidia GPUs. OctoML, the startup Ceze runs, is working on a bunch of gnarly technical fixes that it hopes will make it easier for AI developers to use any GPU with different frameworks and platforms in the future, including AWS's offerings. Second, once you can efficiently connect to another AI software stack, that alternative platform has to be at least as good as CUDA. "And the alternatives are just not there yet," Etzioni said. AMD may be the closest to Nvidia at the moment, according to Etzioni. That chipmaker has ROCm, a software platform that gives AI developers access to popular languages, libraries and tools. He didn't mention AWS or Trainium, though.

You Searched For "shortcuts gpt" and got 6 results

You can now get ChatGPT on your iPhone through this shortcut - here’s how

Rahul Verma

The GPT college classroom is almost here, but professors say AI isn't the real threat

Kai Xiang Teo

OpenAI's ChatGPT can actually make workers perform worse, a new study found

Aaron Mok

I've been building popular tech products like Tome since I became my family's breadwinner at 15 years old. I've learned you never know how people are going to use your creation.

Lakshmi Varanasi

Search engine for your online accounts — Here's how to find anything from your Gmail, Slack, Messenger and similar at once

Rahul Verma

Amazon dominated the first cloud era. The AI boom has kicked off Cloud 2.0, and the company doesn't have a head start this time.

Alistair Barr