The chipmaker's co-founder put into context just how much AI needs his company's tech.

Nvidia's latest earnings report shattered analysts' expectations and sent the AI chipmaker's stock surging to more than $1,000 a share for the first time, as the Taiwan-based company reported first-quarter earnings Wednesday. Nvidia reported quarterly revenue of $26 billion, easily exceeding the average estimate of analysts, pegged at $24.65 billion. The company's chips have become the infrastructural key for startups racing to build large language models (LLMS). Which makes them golden. The stock's run past $1,000 a share pushed its market value to about $2.5 trillion, compared with $745 billion at this time last year. Nvidia and its shareholders are beneficiaries of the AI explosion. On Wednesday's earnings call, CEO Jensen Huang made clear the broader industry's dependence on Nvidia's processors, as demand for AI data centers grows at a rapid clip. Basically, the company is racing to meet almost exponential demand. Huang mentioned OpenAI's flagship products, ChatGPT and GPT-4o, Google's Gemini chatbot, and other big LLM developers such as Anthropic, as feeding the influx of companies piling into the expensive and energy-intensive field of generative AI. But Nvidia is serving more than the big players. There is a crush of startups across multiple fields that depend on the company's chips, from digital design to autonomous vehicles, the CEO said. "There's also a long line of generative AI startups, some 15,000, 20,000 startups that in all different fields from multimedia to digital characters, of course, all kinds of design tool applications -- productivity applications, digital biology, the moving of the AV industry to video, so that they can train end-to-end models, to expand the operating domain of self-driving cars," Huang explained. Huang suggested that Nvidia's meteoric growth will continue. The company is banking on its Blackwell chips, unveiled in March, to lower the cost and energy consumption of building LLMs by 25 percent compared with its current crop of Hopper processors. In an announcement, Nvidia claimed Blackwell will promulgate "a new era in computing." With Blackwell's expected release in Q4, there is concern that chip demand could crater until then, which has analysts worried. Amazon Web Services, which provides cloud computing for a third of the global market, is delaying its order for a batch of Nvidia's Hopper processors, and will instead wait for the release of Blackwell, the Financial Times reported on Tuesday. There is concern that the AWS decision could inform the decision making of other companies, creating a momentary thinning of demand, the report says. But with Nvidia's share price surging to new highs and orders still rolling in, the alarm hasn't manifested in any tangible way just yet. After all, there are 20,000 other startups that haven't changed their thinking so far.

GenAI isn't a threat to labor, it's a threat to lazy software.

Look, I don't have any great love for this current version of genAI. I was one of the inventors of the Automated Insights platform, the first commercially available genAI engine, all the way back in 2010. The idea behind the tech was the same then as it is today -- computers turning data into words (and eventually audio, visuals, and video) to achieve more meaningful input and output. The big difference today is the much broader application of said tech. However, I also believe that the motive behind the tech also remains the same as it was when we started selling genAI over a decade ago. And talking about that motive is where I keep getting into trouble. I don't get dragged often, but I got dragged when I said that this wave of AI was potentially going to end the current, and now dated, flavor of SaaS. This is why corporations and investors are all throwing money at anything with "artificial" or "intelligence" stamped on it. Certain people who should know better screamed at me: "You don't even know what SaaS is!" Yeah. I do. The AI versus SaaS battle is going exactly where I said it was going to go, and here's where it's going next. What I Said I said users don't want another screen. SaaS happened because users didn't want to install and maintain software. There came an inflection point when users decided they could live without a subset of functionality, privacy, security, and processing speed if it meant being able to move to a cloud-based, subscription software model. But the evolution of SaaS didn't stop there. When the mobile-first movement happened, users again decided they could live without screen real estate, and ease of textual input, and they would even take another hit to functionality and processing speed if it meant being able to work in a mobile UI. In both cases, the game of SaaS changed completely, and a lot of the players got left behind. My conclusion: The reason the current players are all panic-investing in genAI is because it represents not just another shift in UI, but the potential to completely change the interpretation of input and output in that it purports to eliminate screens altogether. Once users decide they're willing to live without those screens, genAI exposes SaaS as just a clunky middleman between you and your data: Developers of software and apps will no longer be able to get away with just barfing out aggregated and summarized data back to the user for them to make their own business decisions. When people talk about the uselessness of apps like Google Analytics 4 (GA4), which is actually very useful for people who know how to crunch the data, this is what they're talking about. You know what happens to middlemen during innovation cycles? They get disappeared. What OpenAI (and Google) Are Up To OpenAI released its newest model, GPT-4o, on May 13, and the references to the movie Her -- in which Scarlett Johansson voices a futuristic AI companion -- began immediately. Allegedly, this was by design and somewhat shady. On the surface, the GPT-4o release was a large number of textual, audio, and video improvements. But it doesn't take a data scientist to figure out that the thread between all these improvements was a major upgrade in the interpretation of user input and output. For example, GPT-4o can "look" at a screen of computer code (input) and "summarize" what that code is doing (output). Obvious and somewhat unrelated question: Why does GPT-4o even need a screen to do this? Obvious answer: It doesn't. Anyway, the very next day, May 14, Google held its I/O event and used AI to tell the audience that the company mentioned "AI" in the keynote 120 times. Google is not far behind OpenAI at this point. It announced several new entrants and updates around inputs and outputs, including audio, video(!), and the already hyper-advertised Circle to Search. This is not about chatbots or companions. If you think these advancements aren't coming for the current crop of SaaS players ... well, OK then, let's talk about what SaaS really is. What SaaS Really Is SaaS is not just software in the cloud. SaaS is not just subscription-based software. You can call both of those things SaaS if you want, but you might as well be calling SaaS anything developed after 1990s desktop software. You have to get purposefully vague with the definition of SaaS to stop there. Making the argument that genAI isn't a threat to today's SaaS software empire is akin to making the argument that electric vehicles aren't a threat to today's ICE vehicles because Teslas still have four wheels and an engine. Yet Ford responded to that threat by eliminating production of all, not most, ALL of its ICE sedans. Most other automakers did something similar. Shortsighted? Maybe. But way less shortsighted than blowing off the threat entirely and losing your R&D budget, then your market share, and then your job. Why GenAI Is a SaaS Play and Not an AI Play My problem with genAI is that it's still mostly Gen and not enough AI. It isn't about machines doing the work. It's about machines interpreting the data required to do the work. Left in a vacuum, i.e., without the right prompt, genAI produces mostly random garbage. Even with the right prompt, it sometimes hallucinates and definitely misfires. When I talk about genAI versus SaaS, I'm not just talking about where the processing happens. I'm not talking about how you pay for it. I'm talking about what control over the input and output means to the question of who controls the software. In that argument, software has always been a service, whether the input and output took place on a mainframe, a desktop, or a mobile app. Screens are about input and output. No one ever controlled the keyboard. No one ever controlled the graphical display. OpenAI and Google? They aim to control both input and output. Call it the "interpretation layer." Whoever can best interpret and simplify the question being asked and the answer being produced ... wins. That's a massive threat to what we're calling SaaS today. EXPERT OPINION BY JOE PROCOPIO, FOUNDER, TEACHINGSTARTUP.COM @JPROCO

How to Use AI to Help You Raise Prices and Increase Customer Purchase Frequency

Would an investment in generative AI make your business better off? The answer depends on how much you invest and whether that capital produces a generative AI application that helps your company grow faster. After putting the finishing touches this week on my soon-to-be-published book, Brain Rush, I can say that not a single company I know of has reaped such rewards from their investment in generative AI. My book profiles many companies experimenting with AI chatbots and some of them aspire to achieve such favorable outcomes. You could think of generative AI applications as a pyramid. Here are its three levels. Let employees experiment At the base of the pyramid -- where most companies are trying to invest -- employees get access to an AI chatbot with guardrails to keep them from releasing proprietary company information or inadvertently using the chatbot to compromise the company's reputation. Employees experiment with the chatbots to help them overcome the creator's block they might feel as they compose emails, create marketing copy, or even produce presentations. It is unclear whether the benefits of these generative AI applications will exceed the cost of providing them. Increase functional productivity A far smaller number of generative AI applications aim to increase the productivity of business functions such as writing code or providing customer service. My research suggests such applications have the potential to increase productivity by well over 10 percent, according to companies I have interviewed. I do not know whether the productivity increases from these applications exceed the costs of building and operating them. Moreover, nothing keeps your competitors from building their own such applications -- hence they might not give your company a competitive advantage. Spur revenue growth Based on my research, very few companies are using generative AI to grow faster. The good news is that if such generative AI applications deliver faster growth, their payoff is likely to be substantial. The bad news is that few generative AI applications are so valuable that customers are happy to pay more money for them. The one example that could potentially meet this test is a service from Bullhorn aimed at enabling its customers -- temporary worker placement agencies -- to grow faster by using best practices to make their recruiters more effective at placing workers who succeed. If your company can build a generative AI application that helps your customers grow faster, they should be willing to pay for it. If they do pay, your company will be more likely to earn a return on its generative AI investment. vcita's BizAI helps its small-business customers make more money vcita, a Bellevue, Washington-based provider of software for SMBs that was founded in 2011, offers a generative AI application that enables its customers -- businesses with between five and 50 employees -- to make more money. Although customers like the service -- called BizAI -- vcita is not charging for it. In a May 16 interview, Itzik Levy, the company's founder and CEO, told me he started a cybersecurity company and sold it to Microsoft -- working there for about two years before leaving. Since starting vcita, the company has reached about 200 employees and raised over $33 million, according to PitchBook. "I started vcita to provide software for small and medium-size businesses," Levy says. "Our customers are SMBs with one to five people -- we consider our customers with 30 to 50 people to be big. We have 100,000 paying customers, 70 percent of which are in the U.S." vcita serves SMBs in many industries -- including lawyers, doctors, barbers, realtors, and psychologists. "We started out providing scheduling software for web sites," he said. "We have since broadened our services to include money, compliance, customer relationship management, marketing, and billing." vcita recently launched BizAI -- a free McKinsey-like service that helps SMBs make more money. "By uploading 2,000 words about our clients' business into ChatGPT or Gemini, we can make the AI chatbot much more valuable to them," Levy told me. "For example, one of our clients is a child psychologist who treats patients with ADHD. She asked our BizAI service what price she should charge for her services. Because she has 25 years of experience and is working in a high-demand field, BizAI explained why she should charge at the high end of the typical range -- between $150 and $300 --for a 45-minute session," he added. Should customers pay for such a service? If the company is helping its customers to make more money, they should be willing to share in the gains. If vcita did pay, BizAI could be at the peak of my pyramid of generative AI applications. EXPERT OPINION BY PETER COHAN, FOUNDER, PETER S. COHAN & ASSOCIATES @PETERCOHAN

The OpenAI CEO says generative AI is a once-in-a-generation 'platform shift,' and entrepreneurs need to get on board.

OpenAI CEO Sam Altman has a simple message for entrepreneurs when it comes to artificial intelligence: Now is the time. During an onstage conversation with Microsoft CTO Kevin Scott at the company's 2024 Build conference, an event for developers, engineers, and entrepreneurs, Altman said that right now is probably the most exciting time to be building a startup since either the rise of mobile computing or the dawn of the internet, referring to such moments as "platform shifts." These platform shifts present huge opportunities to "build something new and change the landscape ... and this looks like it's really, truly a platform shift," Altman said, referring to the transformative power of generative AI. Since platform shifts like the ones Altman was describing don't come around very often, he said his biggest piece of advice to entrepreneurs is to take advantage of the moment. "This is not the time to delay what you were planning to do or wait for the next thing," he said. "This is a special moment." Altman's second piece of advice was to not get sucked into the belief that simply integrating AI will be enough to make your company a success. New technology like genAI, he says, "doesn't get you out of the hard work of building a great product, or a great company, or a great service." AI alone will be a great enabler to help businesses get to that next level, he said, "but it does not automatically break the rules of business." At the end of the day, you still have to figure out how to build enduring value for your company, which Altman said is easy to lose sight of in the excitement of the gold rush. Speaking of "breaking the rules of business," Altman, whom Scott referred to as "one of the busiest people on the planet," said he'd been having a "wild week," likely referring to OpenAI's demo of its next-generation voice mode, and the resulting backlash from actress Scarlett Johansson, who threatened legal action against the company over similarities between her voice and one of OpenAI's digital voices. It may have also referred to his public apology on Saturday, after Vox reported that departing OpenAI employees were forced to either sign an NDA or risk forfeiting their vested equity in the company.

IS THIS THE WEEK MICROSOFT UNVEILS ITS AI-BOOSTED PCs?

Microsoft has a special event in Seattle, kicking off its regular Build conference, an event aimed at developers and coders, and the tech world is abuzz with rumors about what the software giant is expected to reveal. Microsoft has relentlessly touted its all-aboard status on the AI bandwagon, publicizing how it has injected artificial intelligence smarts into every part of the Windows experience, from keyboard buttons through to business software. So on Tuesday, it's all but certain that "AI" will be the most-uttered word. And since your office computers are most likely powered by Windows, this means tomorrow's event will most likely show the way you and your co-workers will be using PCs soon. AI in Microsoft hardware Microsoft's been touting the future "AI PC" concept for a while, in concert with its long-term partner Intel, whose "Intel Inside" stickers have adorned billions of PC cases for decades, so it's reasonable to expect the software giant will have more to say about this concept. Microsoft's pitch borrows more than a few ideas from Apple, which has quietly been building specialty AI processing systems into its custom-made processors for Macs, iPhones, and iPads for several years and then tightly integrating these AI processors with what its software can do. Technically, an AI PC will require the computer's motherboard to have a special "neural processing unit" built in. This "NPU" is a tiny cluster of processor power designed specially for the kind of math that makes AI systems like recognition work extra-fast. We can expect Microsoft to tout the promise of AI-equipped PCs, in an effort to tempt regular users as well as business buyers to purchase them in the near future. Until now, Microsoft's been quiet about what exactly they can do, though Intel has indicated that tricks like speech recognition (to boost digital assistant systems), real-time image processing to polish up video calls, and other features, like auto-recognizing features in photos to help you edit them, could feature prominently. Microsoft will likely spend a lot of time promoting its own Surface laptops and tablet devices, as well as putting its Windows operating system on so-called Arm chips, according to tech news site Windows Central. These chips are a much more power-efficient processor rival to the kind of silicon chip Intel and others make for typical PCs. Microsoft's Surface devices have long been praised for their "premium" arty design, in contrast to traditionally clunky PCs, and the new crop will likely be pitched as rivals to Apple's power-efficient MacBooks and powerful iPad Pros, and a way for Microsoft to show off its AI systems on hardware optimized for it. AI in Microsoft software Microsoft-watchers at Windows Central report that the company's "AI Explorer" may be one of the big stars of the event, with a feature named "Recall" as the most dramatic demonstration of how AI will change everyone's computing experience. Recall relies on the neural processing unit running all the time, carefully keeping an eye on what you've been working on on your PC desktop. In a busy work session, perhaps with multiple documents and spreadsheets and so on all open at once, this could help you find that one tricky thing you've lost track of, via a natural-language search, or come up with suggestions about text to write, charts to build for data, and so on--similar to some of the real-time work-watching powers OpenAI revealed last week with its new ChatGPT GPT-4o model. Recall will also let you rewind your work session back to a specific moment, perhaps so that you can change your mind about a particular decision. The concept sounds like it really may save people time at work, though there's no word on whether it'll solve the classic "Windows crashed before you could save your work" error that has generated trillions of curse words shouted across offices for decades, all without any AI assistance. AI critics will worry about the security aspects of this technology, especially against the background of security worries around Microsoft's flagship Copilot AI service. But if Microsoft is savvy about the privacy risks of its users, Recall may work mostly offline, sending no AI data into the cloud where it could leak out. In fact, Microsoft is expected to show off on-device AI processing tomorrow too, with some Copilot capabilities that are currently processed in the cloud being carried out on the PCs' own chips. BY KIT EATON @KITEATON

AS THE AI GOLD RUSH BARRELS FORWARD, VCs EXPLAIN WHAT KIND OF AI STARTUPS THEY ACTUALLY WANT TO INVEST IN

The venture capitalist Gregg Hill spends a good amount of his working hours sussing out which companies are building AI products with staying power, versus those merely looking to cash in on hype. "The majority of companies are incorporating AI into their pitch decks," says Hill, a co-founder and managing partner of Parkway Ventures. "Every startup has dot AI" in their website url, he says, reflecting on the growth of the generative AI landscape. The technology that can generate text, video, audio, and more via written prompts has become entrenched in everyday business lingo and the cultural zeitgeist since OpenAI released ChatGPT in late 2022. It's kept investors busy wading through tremendous amounts of noise, as others make bets seemingly for the hell of it. "A lot of VCs, no offense to them, they're just getting into AI because it's the trend," Hill says. Though it now seems all-consuming, the AI revolution came just as another tech gold rush--characterized by web3 and crypto companies--was losing its luster. Startups and their financiers are now dealing with the aftermath of the industry's record funding boom, which began during the pandemic and wound down in 2022. Compared with the banner year of 2021, when total funding to U.S. startups reached $329 billion, things are much drier now: AI is one of a few categories netting investment anymore, alongside health care and biotech. As a result, startups are piling into the space and barraging investors with AI-enabled this and AI-powered that. Many investors have fallen into the FOMO trap, Nagraj Kashyap, general partner at Touring Capital, explains to Inc. Fear of missing out is spreading through the VC landscape, giving Kashyap flashbacks to the flusher days, when interest rates were rock bottom and money was free-flowing. "That whole era of companies getting priced at valuations that were not defendable ... And you would expect the investor community to learn from that. But we don't see that learning happening," he explains. It may seem like the AI hype train is destined to crash, as a few core giants such as OpenAI, Anthropic, and Cohere absorb funding from tech giants, and compete to build the best large language models (LLMs) in the game. But VCs who spoke to Inc. explain that there is massive promise for smaller players--and the majority of it lies in tools developed by companies that can cater to specific business needs. "Is there a real world problem that is not 'Hey, I can't generate text fast enough'?" asks Amias Gerety, a partner at the fintech investor QED. For founders, Gerety's question could prove instructive. Investors are looking to AI startups that offer solutions for businesses, or provide a modern makeover for technologies that haven't evolved in decades. A lot of the time, the applications are quite niche, even bordering on the mundane. Gerety points to Ocrolus, one of QED's portfolio companies, as an AI startup that has recognized and cornered a particular need: The company's tools extract data from pay stubs and bank statements to help lenders make better decisions. An AI tool for financial underwriting won't carry the mainstream resonance or shock value of a text-to-video tool like OpenAI's Sora. But less flashy, behind-the-scenes tech is what many VCs believe will eventually reshape the working world in AI's image. Kashyap mentions Netradyne, an AI-powered driving monitor that fleet vehicles use with the goal of fewer accidents and traffic violations in mind. "In real time, [Netradyne is] figuring out whether a truck is basically braking too hard, is almost tailgating, is going through stop signs without stopping," he says. As far as moonshot investments are concerned, Hill of Parkway isn't backing the makers of LLMs. But the ambition is still there: Parkway was a lead investor in the $23 million Series A fundraise of Oxos, a medical technology company building what it touts as a "radiology department in a box." Its device, a handheld X-ray machine called AiLARA, automates the appropriate amount of radiation necessary to achieve correct results. Anyone pitching VCs on an AI startup needs to consider how their tools can cater to a specific niche. "Is there a sales process that is broken and could be fixed with better payments? Is there a business interaction where you could make it easier for business X and business Y to coordinate?" Gerety asks, summarizing the kind of pointed questions founders should be asking. Of course, there is always the possibility of a bigger incumbent gobbling up whatever innovation a startup creates. Which is what makes the challenge so daunting. The ultimate sweet spot, Gerety says, is when "it's not just hard for your competitors to copy you, but your product gets better relative to the competition as you grow." The use of enterprise AI tools is already surging. A 2023 survey from workplace productivity tool Asana found that 36 percent of U.S. workers are already consulting AI tools at work at least once a week. The demand is there--and so is the money, if you've got the right specialty.

SOFTBANK IS PIVOTING TO AI. IS IT ALREADY TOO LATE?

SoftBank, the venture capital firm that came to epitomize the hypergrowth era of startups, is--like many players in tech--pivoting to AI. The Japanese investment giant has been writing down investments and selling shares in many of its portfolio companies, presumably to cut losses that have piled up over the past few years, according to various reports in The Wall Street Journal and Bloomberg. The timing may be fortuitous, as the refocus is fueling SoftBank's plan to hitch itself to the generative AI rocket ship. The Tokyo-based VC firm saw the net value of its assets climb to 27.8 trillion yen, or $177 billion, at the end of the past fiscal year, Yoshimitsu Goto, SoftBank's chief financial officer, said on an earnings call Monday. The surge was predicated on the performance of its marquee asset in AI--the chipmaker Arm, which has surged since it reentered the public markets in February. The release of its AI chips is hotly anticipated, though it won't release them until next year. The chipmaker's shares are now trading at $115, up 90 percent from last September. Now, SoftBank wants Arm to lead its efforts in the AI era. "Arm and portfolio companies should create a new ecosystem going forward. That's our expectation and that is reflected in this portfolio," Goto said on the call. Softbank seems to be putting a lot of eggs in one Arm-shaped basket. "Everything is all about edge in investing. You want to figure out what's your edge," Steven Kaplan, a professor of entrepreneurship and finance at the University of Chicago, explains to Inc. "And now the question is, do they have any edge in AI? And it's not obvious." SoftBank didn't return Inc.'s request for comment. Arm was founded in 1990, as a venture between Acorn Computers, a now defunct British tech company, and Apple. It originally operated out of a turkey barn in Cambridgeshire, England, according to the company's official history, and became a key player in the development of mobile phone technology. Historically, Arm has made blueprints for chip designs, and not physical chips, like Qualcomm or Nvidia. It plans to start producing AI chips next year, however, financed by SoftBank, a report in Nikkei Asia revealed Sunday. Arm's potential as an AI chip designer became apparent during the AI boom, as demand for computational power grew among the firms employing the tools. "Arm is riding on the coattails of demand for Nvidia's technology, particularly its data center systems," Susannah Streeter, head of money and markets at Hargreaves Lansdown, told Reuters in February. SoftBank is piling into the generative AI gold rush following years of losses. In 2023, SoftBank reported net losses of 227 billion yen, or $1.46 billion, for the fiscal year ending in March. That was an improvement from 2022, when the firm was out 970 billion yen--or roughly $6.2 billion--as its flagship Vision Fund, known for splashing on WeWork, Uber, and FTX, posted losses of $32 billion. The firm's P&L statement is inching from red to black, however, as SoftBank netted a $1.5 billion profit in Q1, Goto told investors. Speaking of the previous year's losses, he explained: "For us, it was not big, but if you look at the change from the past year, net income improved." SoftBank invested $32 billion into Arm in 2016, taking the U.K. company private for a time. Goto praised Arm's potential to provide highly sought-after computing infrastructure for the AI boom, noting it could service a new "ecosystem" of AI startups bankrolled by SoftBank. Goto mentioned GreenBox, which makes AI-enabled warehouse technology, and Berkshire Grey, a robotics startup, as marquee companies in SoftBank's new AI portfolio, worth around $5 billion. The investment giant's founder and CEO, Masayoshi Son, has lauded the future of AI, saying last October that AI capable of surpassing the total sum of human knowledge--or general artificial intelligence--is only a decade away. SoftBank helped catalyze the VC boom in the days preceding the pandemic. Through its Vision Funds 1 and 2, SoftBank made two separate $3 billion investments into the now bankrupt WeWork, netting a paltry return of $120 million, PitchBook data shows. "In 2021, they helped push the market by putting lots of money to work at crazy valuations. And it actually helped ruin some companies," Kaplan says. There are indications that years of losses, predicated on some of the era's blockbuster failures, has put SoftBank in a bind. "I don't have any insider information. But I would not be surprised to learn later that SoftBank has a liquidity crunch," Ilya Strebulaev, a finance professor and economist at Stanford, tells Inc. Of course, jumping in after the initial wave just might be SoftBank's preferred mode of operation. WeWork was around for six years before SoftBank invested. "I don't believe SoftBank's ever been a first mover," Strebulaev explains.

FIGHTING FIRE FOR FIRE: 3 STRATEGIES FOR PROTECTING YOUR CONTENT FROM AI THEFT

The bad news is that AI is stealing our content. If you're a content creator--a writer, videographer, artist, musician, or similar--AI may have already repurposed your content. That's why Getty Images, Universal Music Group, George R.R. Martin, Sarah Silverman, the New York Times, and more are suing AI developers for copyright infringement. At the same time, in the business world, those who don't use AI risk falling behind or producing work at a rate that's too slow to keep up with new expectations and demand. So, the paradox for creators is clear. And AI isn't going anywhere. Language learning models like ChatGPT and Claude have become ubiquitous fixtures of creative workflows. AI image generation tools like Dall-E and Midjourney have been used in ads by big brands like Burger King and Progressive, and a search for "AI music" reveals pages of tools anyone can use to create generative AI music. In fact, AI use has become so prevalent that experts forecast global market revenue of AI usage in marketing alone to reach $36 billion this year. AI is disrupting industries across the board. In a blog post outlining how AI will change how we use computers, Bill Gates explained that individuals and companies will eventually have personal "AI agents" capable of performing just about every task imaginable, from scheduling appointments to designing customized apps. So how about that good news I mentioned? Well, the good news is that you can not only use AI to (ethically) enhance your workflow--you can also use it to fight back. Strategy 1: Create an AI Agent of Yourself One of the most helpful things about language learning models is that you can train most of them to mirror your specific voice. Projects like Coachvox AI, founded by Jodie Cook, allow creators, authors, coaches, and consultants to train AI models with their original content, preserving their unique style and perspective and protecting their IP. These AI avatars can interact with your audience, providing mentoring and guidance based on your frameworks and insights. You can also use your AI avatar to bounce ideas off of "yourself" and speed up your creative process. Creators can even monetize by charging a monthly fee for access to the personalized guidance and support their AI avatars provide. This is a fantastic, ethical way to scale impact without sacrificing quality or brand authenticity. Strategy 2: Deploy AI Poison and Code Blocks Computer scientists Ben Zhao and Heather Zheng at the University of Chicago recently launched two tools, Glaze and Nightshade, that protect visual creative work from being used without permission by AI models. Both AI tools "poison" AI outputs by introducing changes to digital content that are invisible to humans, but disruptive to AI training and scraping processes. Glaze is intended to protect individual artists' styles from being mimicked by AI. This tool alters some of an image's pixels in ways the human eye doesn't notice while fundamentally destroying what an AI can understand about it. So, a "Glazed" image of a cartoon puppy could cause AI to "see" a piece of cake in the style of Van Gogh. Nightshade takes AI poisoning one step further. Where Glaze takes a "defensive" approach, Nightshade is considered an "offensive" tool that actively sabotages what AI models will output for users in future sessions. Like Glaze, Nightshade adds subtle changes to pixels in artwork that corrupt AI model training data. This pollution causes affected AIs to learn wildly incorrect patterns and produce unpredictable, incorrect results that don't align with user prompts. And while we don't yet have "poison" for written content scraped by LLMs, you can add code blocks on your website--like the ones implemented by Tony Stubblebine, CEO of Medium--to block content from being visible to AI crawlers. Strategy 3: Build and Share on Dark Social More than ever, people consume publicly but share privately through text messages, Discord channels, WhatsApp chats, emails, and more. Basically, if a platform has DMs, you can expect that people are sharing content through them. This creates a "dark social" realm where marketers and trackers--and AI--can't see what's being shared. While this might not be a comprehensive solution (individual users could potentially still copy and paste your content into various AIs), it is one way to fight against the kind of internet-search-AI-scraping models like ChatGPT and Google Gemini have the power to perform. And here's something else: AI can't access information from password-protected platforms. So, if you publish your content on a site that requires authentication to access--like Patreon, private Facebook groups, and similar--AI can't see it (at least, for now). So, while AI may pose challenges to content creators, creators are far from powerless. By using AI tools to enhance your own creative processes, deploying emerging technologies to protect your work, and being intentional about where and how you share, creators and innovators can harness the power of AI rather than be steamrolled by it. EXPERT OPINION BY SHAMA HYDER, FOUNDER AND CEO, ZEN MEDIA @SHAMA

MICROSOFT'S AI COPILOT'S ENTRY-LEVEL JOB: DOING WORK YOU'D GIVE TO AN INTERN

Despite the tech evangelists constant chorus of praise for the amazing powers of current-generation AI, it's pretty clear that it's not going to be replacing CEOs or out-innovating entrepreneurs building startups anytime soon. AIs are just not that sophisticated. One thing the current AI technology can do well, though, is doing simple format, repetitive tasks. That's how one company appears to be really getting its money's worth out of Microsoft's Copilot AI. It's using it to automate many of the typical, boring bits of office work that might be given to the office noob or the intern. Technology news site ZDNet reports Amadeus, a multinational tech company that sells software systems to the travel industry, has embraced Copilot. The simplest thing it's using Microsoft's advanced AI for may seem blindingly obvious: It's automatically transcribing notes from meetings. In an interview with website Skift.com, Frederick Ros, the company's head of digital services explained that "There is no organization in the world that is really super good at taking notes of meetings." More importantly, he added "it's always difficult to have somebody focusing on what is said and taking notes instead of being part of the discussion." That does sound like many typical business meetings, and it may be why sometimes note taking is left to a staff member who isn't actually expected to take part in the chats, like an intern. But Copilot can apparently do it all, including accounting for different employees' accents and absorbing the company's specific business acronyms. The company also said it was using Copilot to summarize "long threads of chat between co-workers, which is particularly helpful for those who join a chat late." That's something that workplace digital chat app Slack has built into its software recently for the same reason. Amadeus is also using the AI to help employees draft emails -- which they're finding is especially useful "for non-native English speakers" -- and to help staff make better searches when they're looking for relevant information online or even analyzing the company's own data. The company seems to have fully embraced AI, and Ros explained that it's pushing to get staff to use Copilot daily so they "will have this deep understanding of how it works and what they can bring, and it will probably spark some ideas in their mind." That last idea chimes well with recent news from drugmaker Moderna, which also deeply integrated OpenAI's tech into its entire business structure. The company's CEO recently surprised people by saying he expects his staff to be consulting ChatGPT at least 20 times per day in order to streamline their work. It's glib to say that Amadeus is using Copilot to do work an intern would typically do, but the saved seconds from using an AI to, say, start off an email, all add up. Saved time can either contribute to a company's bottom line (since you may actually need fewer staff) or allow staff to get on with more meaningful tasks. And in this way, maybe that's actually the superpower of current-gen chatbot tech: it's all saving small companies with few employess critical time to actually do real work. The one issue with all this, of course, is that AIs can't be 100 percent trusted right now. Just as a junior staff member could make a slip-up when taking notes during a complex finance meeting, AIs are not foolproof, and current-gen chatbots can misunderstand instructions and even hallucinate totally fake responses to queries. So if you're being thorough while using AI to streamline some of your business processes, someone still has to go through and make sure it's giving the right answers -- lest you end up with bad notes or even in expensive legal trouble because of something an AI sent in an email. That sounds like a job for an intern. BY KIT EATON @KITEATON.

WHY CYBERSECURITY IS MORE IMPORTANT THAN EVER

In today's interconnected world, where data flows freely and technology permeates every aspect of our lives, cybersecurity stands as a paramount concern. From personal information to critical infrastructure, the digital landscape is fraught with vulnerabilities that can be exploited by malicious actors. As our reliance on technology deepens, the importance of robust cybersecurity measures cannot be overstated. Cybersecurity encompasses a broad range of practices and technologies aimed at protecting digital systems, networks, and data from unauthorized access, exploitation, and disruption. Its significance extends beyond individual privacy concerns to encompass national security, economic stability, and societal well-being. One of the most pressing cybersecurity challenges is the protection of personal data. With the proliferation of social media, online shopping, and digital services, individuals routinely share sensitive information online. From financial details to personal preferences, this data is a prime target for cybercriminals seeking to commit identity theft, financial fraud, or other nefarious activities. Strong encryption, secure authentication mechanisms, and data protection regulations such as the General Data Protection Regulation in the European Union are crucial for safeguarding personal privacy in the digital realm. Businesses and organizations face their own set of cybersecurity challenges, particularly as they increasingly rely on digital infrastructure for day-to-day operations. Cyberattacks targeting corporations can result in financial losses, reputational damage, and disruption of services. The rise of ransomware attacks, where cybercriminals encrypt an organization's data and demand payment for its release, has become a particularly pervasive threat. To mitigate these risks, companies must implement robust cybersecurity protocols, conduct regular security assessments, and invest in employee training to foster a culture of security awareness. Critical infrastructure, including power grids, transportation systems, and health care facilities, represents another prime target for cyberthreats. A successful cyberattack on critical infrastructure can have far-reaching consequences, disrupting essential services and potentially endangering lives. As these systems become increasingly interconnected and reliant on digital technologies, the need for stringent cybersecurity measures becomes even more pronounced. Governments play a crucial role in this regard, establishing regulatory frameworks, investing in cybersecurity infrastructure, and fostering collaboration between public and private sectors to enhance resilience against cyber threats. The proliferation of the internet of things further complicates the cybersecurity landscape. From smart thermostats to connected cars, IoT devices introduce new entry points for cyberattacks, often with insufficient built-in security measures. Compromised IoT devices can be harnessed to launch large-scale cyberattacks, such as distributed denial of service attacks, which overwhelm targeted systems with a flood of traffic. Securing IoT devices requires a concerted effort from manufacturers, consumers, and regulatory bodies to enforce security standards and ensure the integrity of these interconnected systems. In response to the evolving cybersecurity threat landscape, organizations and governments are increasingly turning to advanced technologies such as artificial intelligence and machine learning to bolster their defenses. AI-powered cybersecurity solutions can analyze vast amounts of data in real time, identifying and mitigating potential threats more effectively than traditional approaches. However, as with any technology, AI also poses its own security risks, including the potential for adversarial attacks and unintended consequences. Ultimately, cybersecurity is a multifaceted endeavor that requires a comprehensive approach encompassing technological innovation, regulatory oversight, and collaborative efforts across sectors. As we navigate an increasingly digitized world, the need to prioritize cybersecurity has never been more critical. By investing in robust cybersecurity measures, fostering a culture of security awareness, and promoting collaboration between stakeholders, we can build a safer and more resilient digital ecosystem for generations to come. EXPERT OPINION BY GREG TOMCHICK, PARTNER AND CEO, VALOR CYBERSECURITY @GTOMCHICK

HOW YOU CAN USE AI TO STREAMLINE YOUR SOCIAL MEDIA MANAGEMENT

Having a solid tech stack is more important than ever in this era of AI. The mainstream adoption of AI requires brands to streamline and improve their internal processes. Brands are already using AI to great effect in social media management. By understanding how other brands use AI to empower their teams and improve social media practices, you can find (and create) opportunities for improvement in your own work. 1. Generate post content. When considering AI social media applications, the first thing that probably comes to mind is creating content using generative AI. Despite its current capabilities, you can't assume an AI should be able to produce content that matches your brand voice and appeals to your audience on its own. AI has simplified content creation for many brands. For example, HubSpot describes how brands can quickly generate high-impact social shares by feeding the copy from their own data-driven report into an AI tool. AI can similarly repurpose blog content and other resources to produce engaging and informative posts. 2. Content ideation. Obviously, you can't fully rely on your pre-existing content to fill out your social media calendar. Fortunately, AI tools make creating new content ideas easy. As described by Wired, providing AI tools with specific prompts (such as asking for lists of content ideas on a particular topic) can provide a lengthy list of ideas you can use for your own content. Some social media-focused AI tools go even further, generating multiple post ideas with content suggestions based on general ideas for your niche. 3. Understanding audience sentiment. Social media posts need to be monitored constantly, but this can take a lot of effort. To refine your content strategies, you must understand why a post resonates or fails to engage your target audience. Edward Rivers, CEO of Comb Insights, recently explained how his company created a proprietary AI that could scan comments on Facebook and Instagram posts to show users the percentage of positive, neutral, and negative comments on each post. By assigning posts an overall positivity score and showing commonly used words in comments, brands can quickly assess how well a post performed, and why. 4. Optimize social media advertising campaigns. AI can be a powerful tool in paid social media advertising campaigns. For example, content creation tools can quickly generate multiple ad copy variants, helping you quickly create unique ads for different audiences. Some brands have even created AI influencers to better connect with their audiences, though the calls for transparency regarding these campaigns indicate brands should proceed with caution. However, AI's full potential in social media advertising can be felt on the strategic side. Predictive analytics and real-time results analysis can improve advertising strategies and audience targeting. Artificial intelligence can help brands maximize their results by optimizing real-time targeting and bidding. 5. Streamline scheduling. Scheduling your social media content can be a surprisingly time-consuming manual task, even when you have basic scheduling software to assist you. AI tools can help brands take scheduling to the next level, with abilities such as bulk scheduling or locking in post categories in a weekly calendar so new posts can automatically be assigned to the right date and time. AI analytics can then track your posts' engagement, helping you identify the days and times when your audience will most likely engage. With a few clicks, you can adjust your scheduling to expand your reach. AI can also facilitate A/B testing to see how different posting strategies compare in terms of overall performance. Maximize your social media potential with AI. AI isn't going to completely take over your social media team's work. A major appeal of social media is its ability to connect people through digital interactions. In addition, AI can empower your team to take on higher-level work that allows you to connect with your audiences more effectively. EXPERT OPINION BY HEATHER WILDE, CTO, THEDIFFERENCE @HEATHRIEL

AMAZON OFFERS ITS Q AI ASSISTANT TO BUSINESS USERS

Amazon is far more than an online store giant--its global Amazon Web Services (AWS) cloud system provides core IT infrastructure for businesses of almost every size. Now AWS is rolling out a new AI system called Q that embraces the buzziest of buzzy technologies, allowing businesses to "transform the way their employees get work done." The company blog explains that Amazon Q is "the most capable generative artificial intelligence (AI)-powered assistant for accelerating software development and leveraging companies' internal data." That's sales pitch lingo at work--but what Q offers is actually quite powerful, and might live up to the hype. Since this sort of AI assistant is all about enabling companies to do more with less, benefits to smaller enterprises with smaller teams with less diverse experience could have a greater impact that it might at larger companies. Q's capabilities appear to split into three major parts. The first is all about coding--a vital skill for many businesses, even if their core offerings aren't based on technology. That's because nearly everything in our world is digital, and that code is its working language. Amazon's blog post says developers claim they spend less than 30 percent of their time on actual coding, while the rest of the day is spent on "tedious and repetitive tasks" and managing IT infrastructure matters. This is "coding muck," Amazon says, and it obviously hits company bottom lines. Q takes aim at this muck, harnessing AI tools that can perform tedious tasks like quickly debugging already-written code automatically. Amazon follows industry trends here--writing code is a key ability of many existing chatbot systems, and Microsoft is even tying AI systems into its social media-like code repository GitHub. Current-generation AI systems also excel at analyzing and summarizing huge amounts of data, something Q includes to help business users. The company's press release includes a slide showing Q being used as part of Amazon's "business intelligence" Quicksight system, quickly summarizing an imaginary company's sales and profit data. Done by an employee, this is a tedious, time-consuming task that requires specific expertise. A small company can use Q to bolster its expertise and save lots of time, Amazon says. Customized AIs can be notably beneficial to individual companies--either because a custom-trained AI can focus on business-specific needs. A small business can use it to create, say, a customer-facing chatbot that understands specifics about its products and business operations, and explain them quickly. Q touts its ability to do this, and also includes a sort of "no-code" approach, meaning any employee can ask it to create a custom AI system just using straightforward language like: "Build me an onboarding app for our new hire Sarah." Amazon certainly isn't the first company to offer all this sort of business-centric AI tech. Microsoft, OpenAI and many others offer some of the same services. But the tech giant's advantage here is its ready-made customer base: many companies already use Amazon's cloud systems making it easy to pitch its AI systems to them as well. Amazon's hawking its Q AI very hard, and notes that since its unveiling, several companies are testing it. Based on this it says Q can help "employees become more than 80 percent more productive at their jobs." Hopefully this means that employees are more relaxed and, thanks to AI, have a wider skill set with which to tackle tasks without being overworked. There is some evidence that using AI harms staff well-being, and previous studies show some employees are wary of embracing AI tech. Their biggest suspicion is they're merely training an AI that will eventually replace them, and there's some truth to this. On that note, Amazon's press release says with the confidence of a tech evangelist that "new features we're planning on introducing in the future," employee productivity "will only continue to grow." BY KIT EATON @KITEATON

MODERNA'S CEO SAYS STAFF SHOULD CONSULT CHATGPT 20 TIMES A DAY

OpenAI is seemingly everywhere now--its ChatGPT system is in the vanguard of bringing AI to the masses. It's also laced through the workings of pharma giant Moderna, thanks to a deal weaving OpenAI tech deeply into the fabric of the company. So much so that Moderna CEO Stéphane Bancel says his staff should be making the most of the investment in AI and using it a lot. More than "a lot." Bancel's suggestion: More than 20 times a day. Assuming a typical eight-hour workday, that means Bancel expects his staff to ask OpenAI's chatbots questions at least two, maybe three times every hour--an AI interaction rate that could easily, with some back-of-the-napkin estimates, eat up 10 to 15 minutes of work time each work hour. Maybe more. So what will all that employee-AI interaction do for Moderna? The Wall Street Journal quotes Bancel saying AI is going to be used to "reinvent all of Moderna's business processes, in science, in legal, in manufacturing--everywhere." So far, Moderna's staff have built many different custom GPTs using OpenAI's tech, the WSJ says. These are specially trained versions of the chatbot that are separate from the main ask-it-anything open-access ChatGPT system that most people have tinkered with online. Of the 750 or so custom Moderna GPTs, some are being used to help decide drug doses for clinical trials, and have presumably been trained on proprietary data from previous Moderna trials, while others have more business-specific uses, like helping Moderna deal with government regulators. An official statement from Moderna underlines Bancel's enthusiasm for the technology. He explained that AI is as impactful as the "introduction of the personal computer in the 1980s" which "changed the way we work and live." The goal for such widespread company adoption of AI tech is to support Moderna's "ambitious plan to launch multiple products over the next few years." This gives us a deeper clue as to what Bancel thinks AI can offer for his company: as a multiplier, enhancing the productivity and efficiency of his staff, in any division of the company. The company statement also quotes OpenAI CEO Sam Altman: Moderna's simply "leading the way by empowering all of its employees to use AI to tackle complex problems," he said. The WSJ says Altman also explained that right now, ChatGPT may not be used to advance Moderna's scientific progress too much, though it will be able to tackle "more and more" scientific tasks "eventually." Right now, the best way for Moderna to advance its scientific objectives is "to enhance the productivity of people and accelerate what they can do," Altman said. What can your company do with this AI-embracing approach? Until now, it's been easy to see that content generated by AIs can boost simple business processes like building presentations or preparing marketing material, but Moderna's example shows that by embracing new technologies like custom-trained GPTs, AI could actually be used in almost every part of your business. It's just a question of working out where. If you're hesitant to embrace AI, or your staff themselves are stressed out by the tech's implications or are worried about simply using it, then maybe it's time to take some AI training for you, your management team, and your frontline staff.

3 KEY BUSINESS AREAS CYBERSECURITY FALLS SHORT

Putting your finger on exactly what drives business success is impossible because success is driven by a combination of factors. However, there is one thing that can compromise business resilience and your trajectory--cyber risk. Here are three key areas that drive business success and are at risk without an approach to network security that is designed for modern networks and operating models. 1. Innovation According to the most recent numbers available from the National Science Board, on a global basis, the U.S. leads in research and development with U.S. businesses spending over $608 billion on innovation in 2021--a 12 percent increase from the prior year. And these investments pay off. The top 50 companies in BCG's 2023 Most Innovative Companies report outperform on shareholder return by 3.3 percent per year. However, these days work environments are so diverse and dispersed that R&D teams are working on different clouds or even on-premises. Collaborating via any configuration in the modern network makes it particularly challenging to ensure only the people who are supposed to be working on an R&D project are in fact on the R&D project. This lack of visibility can put your intellectual property and trade secrets at risk. 2. Third-party ecosystem The average organization does business with 11 third parties and each of those third parties has a pathway into your organization--whether through technology integrations or supply chain processes or access into your environment as part of a service they deliver to the business. To mitigate exposure to risk from your third-party ecosystem, verifying that your suppliers have achieved SOC 2 compliance and have implemented a supply chain integrity process and a notification process if their supply chains are compromised are great places to start. Even still, 98 percent of companies do business with at least one third-party partner who has been breached. From a compliance, security, legal, and procurement perspective there are many reasons for concern. 3. Customer relationships The movement to meet customers where they are emerged nearly 15 years ago and now, depending on your industry, customers may expect to connect with you online, virtually, and through social media channels, in addition to in-person, phone, and email. Digital strategies are mostly cloud-based which means you are working with your cloud service providers and technology vendors to enable these services. Still, you are responsibile for protecting the data that flows through these channels and wherever it is stored, which can be overwhelming, particularly in today's multi-cloud environments. Think about cybersecurity differently The traditional approach to network security has been to focus on securing the network as well as possible and then detecting particular, known attacks. But there should never be a quantifiable threat coming from any of these vectors. It's how we collaborate in R&D or interoperate with partners and customers that can create opportunities for compromise. Beyond the known threats to every network, some activities should never occur and the ability to detect the behaviors that are known to be operationally out of bounds is incredibly powerful. Unfortunately, most organizations cannot detect that activity. When users, applications, data, and devices are spread across your multi-cloud and on-premises environment, how do you know what you've got, what it's doing, and what's happening to it? You need comprehensive visibility of all the participants across your environment and the ability to apply policies to enforce behavior that is normal or expected and alert you to activity that is not compliant. For example, when it comes to innovation, there's always the risk of IP being stolen and exfiltrated. R&D teams need to be segmented off from the rest of the organization using zero trust best practices of both identity-based access control and network segmentation to keep unauthorized users from accessing what's being worked on. These same best practices are also essential to have in place to mitigate risk from your third-party ecosystem or your customer-facing touchpoints. And since cloud misconfiguration issues are a major cause of data security breaches, it's also important to validate that your cloud infrastructure is configured and running properly. Rethinking network security to focus on comprehensive real-time observability of the activities of the users, applications, data, and devices across the entire multi-cloud and hybrid environment lets us see when things go awry. We can detect signs of abuse, misuse, or compromise to build resilience and continue on a trajectory to business success. EXPERT OPINION BY MARTIN ROESCH, CEO, NETOGRAPHY @MROESCH