Firefly

John Skogman har en bra börsfredag. Poddprofilen och daytradern är inte bara storägare i fredagens rapportvinnare Duroc. Han har dessutom sålt av nästan allt i Firefly före fredagens rapportfall – aktien som är hans bästa affär i livet: ”Det är den som gjort mig rik, typ – inte miljardärsrik, men traderrik.”

John Skogman har en bra börsfredag. Poddprofilen och daytradern är inte bara storägare i fredagens rapportvinnare Duroc. Han har dessutom sålt av nästan allt i Firefly före fredagens rapportfall – aktien som är hans bästa affär i livet: ”Det är den som gjort mig rik, typ – inte miljardärsrik, men traderrik.”

Fejkporr-skandalen på Westfield high school i New Jersey har skakat om samhället och myndigheterna försöker yrvaket reda ut om porrbilderna kan utgöra ett brott. AI-genererat material har exploderat online – och nischade sajter för manipulerade porrfilmer tjänar pengar på kvinnor som inte har en aning om att deras ansikten utnyttjas. ”Deepfake-porren är bortom all kontroll”, konstaterar techmagasinet Wired. – Vi vet att det finns creepy killar därute, men trodde aldrig att våra klasskompisar skulle kränka oss på det här viset, säger en av de drabbade eleverna på Westfield high school i New Jersey till Wall Street Journal. Tonårspojkar på skolan har använt ett AI-verktyg för att skapa pornografiska bilder och sedan spridit materialet i chattar. Vanliga foton på tjejerna i klassen har genom AI-verktyget förvandlas till kränkande nakenbilder. Upptäckten ledde till stora protester och en rad polisanmälningar. Skolan inledde också en egen utredning och fördömde i starka ordalag elevernas beteende. Polisen har svårt att avgöra om de AI-genererade porrbilderna faktiskt utgör ett brott enligt gällande lagstiftning. New Jersey-senatorn Jon Bramnick har satt en åklagare på att utreda fallet och lovat att stifta en lag som kriminaliserar skapandet och spridandet av porrbilder som genereras med hjälp av artificiell intelligens. – Det här måste klassas som ett allvarligt brott, säger han till Wall Street Journal. Fejkade sexbilder är inget nytt. Photoshop och andra redigeringsverktyg har länge använts för att manipulera och fabricera bilder. Men i dag är det enklare och billigare än någonsin att med hjälpa av AI skapa egna porrbilder med verkliga personer som förlagor. Allt som krävs är en mobiltelefon och en bild på vilken person som helst. Det finns bland annat appar som kan byta ut ansikten i porrfilmer. Ofta är det svårt för det mänskliga ögat att avgöra om en bild är falsk eller äkta. Techmagasinet Wired har undersökt 35 av de största webbsidorna som helt eller delvis riktar in sig på så kallad deepfake-porr. Under årets tre första kvartal laddades totalt 113 000 AI-manipulerade porrfilmer upp på sajterna – en ökning med över 50 procent från i fjol. Forskarna tror att trenden kommer att fortsätta uppåt och Wired konstaterar att ”deepfake-porren är bortom all kontroll”. I Spanien har polisen inlett en utredning efter att nakenbilder på unga flickor – genererade via en app som ”klär av” personen på bilden – spridits på nätet. Regionpolitikern Mariá Guardiola har fördömt den ”vidriga incidenten” och skriver på X att ”digitalt våld mot kvinnor är ett gissel som breder ut sig”. Tidigare i år dömdes en 22-årig man från Long Island i USA för att ha spridit falska porrbilder på sina studiekamrater. Han hade använt bilder från när de nu vuxna kvinnorna var i mellan- och högstadieåldern för att skapa materialet. Mannen dömdes till fängelse i sex månader för sexualbrott mot barn. Flera stora AI-verktyg, som Open AI:s Dall-E och Adobes Firefly, har inbyggda spärrar som hindrar användarna från att skapa pornografiska bilder. Men en uppsjö av andra mindre nogräknade appar erbjuder tjänsterna. Eftersom lagstiftningen snabbt halkar efter teknikutvecklingen är det svårt att hålla bolagen juridiskt ansvariga för händelser som den i New Jersey. Men försök görs. I veckan presenterade USA:s president Joe Biden ett omfattande regelverk för AI, det första någonsin i sitt slag. – Vi har en moralisk, etisk och samhällelig plikt att se till att artificiell intelligens används och utvecklas på ett sätt som skyddar allmänheten från potentiell skada, sa vicepresident Kamala Harris. Presidentens AI-paket innehåller bland annat krav på att märka upp AI-genererat material och regler som förbjuder manipulerade porrbilder på personer som inte gett sitt samtycke. Läs mer

Open AI:s vd Sam Altman räds inte konsekvenserna på arbetsmarknaden som AI-revolutionen leder till. Det sa han på Wall Street Journals årliga konferens Tech Live i veckan: – Varje teknologisk revolution får följder för arbetsmarknaden. Jag är inte det minsta rädd för det. Under mötet samlades världens techledare – från bland annat Meta och Arm – för att diskutera de snabba förändringar som AI medför för näringslivet och politiken. Tidningen listar här några höjdpunkter från samtalen. OpenAI’s Sam Altman, Meta’s Chris Cox and others spoke at WSJ’s annual event. By WSJ Staff

The Wall Street Journal, 18 October 2023 Tech leaders convened on The Wall Street Journal’s annual Tech Live conference this week, where discussions focused on the fast-paced changes wrought by artificial intelligence across business, technology and policy-making. Here are some highlights from interviews. AI has been a central topic this year, as its impact on business and society is hotly debated. Sam Altman, chief executive of OpenAI, said the two things that will matter the most over the next few decades are abundant and inexpensive intelligence, and abundant and cheap energy. OpenAI is working to make ChatGPT cheaper and faster, so that it can be more broadly accessible. “If we can get these two things done in the world then it’s almost, like, difficult to imagine how much else we could do,” he said. Mustafa Suleyman, CEO of Inflection AI, said the race to build AI chatbots is reminiscent of the rush to build websites at the dawn of the internet or apps after the advent of smartphones. “This is really just the beginning of a complete transformation in the way we interact with computers,” said Suleyman, whose company makes a ChatGPT rival called Pi, short for personal intelligence. Meta is also optimistic about the impact of AI. “One of the most profoundly impactful applications in the near term for AI is helping businesses be more effective,” said Meta Platforms Chief Product Officer Chis Cox. Meta last month unveiled its own AI chatbots based on celebrities such as Naomi Osaka and Snoop Dogg. Cox said Meta is making it clear these characters aren’t the real people. “Having products that experiment with what is possible is great, but having anything that doesn’t make clear to people what is going on is a problem,” he said. Consumers are going to gravitate to TikTok, ChatGPT and other applications powered by generative artificial intelligence, instead of using traditional search engines, said Michael Wolf, co-founder and CEO of consulting firm Activate. He predicts that domination within the $100 billion search industry is “up for grabs,” adding that the rise of open-source AI models is paving a pathway for smaller entrants to meaningfully compete with large, established companies. Professionals from physicians to writers have been fearing that AI will entirely replace some jobs. “Every technological revolution affects the job market. I’m not afraid of that at all,” said OpenAI’s Altman. “That’s the way of progress. And we’ll find new and better jobs.” Still, it’s not going to be a seamless process. “The thing that I think we do need to confront as a society is the speed at which this is going to happen,” he added. Adam Wenchel, chief executive of AI company Arthur, took a more sanguine view of the job impact from AI than some other panelists at Tech Live. “These systems are going to roll out over time, very gradually, people are going to adapt to them and it’s going to be OK,” he said. Indeed, companies are still determining how to implement new AI technology. “Even at the highest levels, we’re still trying to figure out what does all of this mean to our business model,” said Vince Marin, chief information officer of law firm Sidley Austin. Charles Sims, chief technology officer at United Talent Agency, said AI makes it more important for people to have generalized skill sets that enable them to adapt as technology replaces specific specialties. “If you’re talking to a college student today, it’s about generalization, it’s about trying to learn as many things as you can,” he said. Elise Smith, CEO of Praxis Labs, said it is critical to involve the next generation of workers in discussions about how to use technology: “They want to be brought in and brought along on the journey,” she said. “They want to be doing the innovation day, the hackathon, where they’re getting to give ideas around how AI can transform their business.” Adobe’s president of digital media business, David Wadhwani, said that despite fears, he sees artificial intelligence as a tool that will boost employment rather than put people out of jobs. Tools like Adobe’s Firefly, which can generate images and logos, allow more people to become creative professionals, he said. “We will have creative professionals being more productive than ever before and more creative professionals in the world,” he said. Arm CEO Rene Haas said the chip company is using artificial intelligence to help in some of the areas where they struggle to hire enough talent, such as with debugging and testing chips. But he said the semiconductor industry faces some challenges in its role powering artificial intelligence. He described a future when energy shortages could constrain AI advancement, and a shortage of talent could limit production of semiconductors. “The kind of people we are looking to hire are hard to find. We are looking for really expert engineers,” Haas said. Investors are weighing whether it’s too late to get into AI. “Most investments in AI today—venture investments—will lose money,” said venture capitalist Vinod Khosla. Khosla, who founded Khosla Ventures almost two decades ago, said AI investing had entered a hype cycle, and only highly disciplined investors will reap the benefits of the transformative technology. The buzzy new technology has generated significant concerns, though. “We’ve got a fierce task ahead of us to figure out what are these downsides and discover, understand them, and build the tools to mitigate them,” said OpenAI Chief Technology Officer Mira Murati. For instance, sometimes chatbots confidently espouse information that doesn’t seem to be justified by its training data. “We’ve made a ton of progress on the hallucination issue with GPT-4, but we’re not where we need to be. We’re sort of on the right track.” Murati said. OpenAI is continuing to use techniques including reinforcement learning with human feedback to reduce the number of times that its model makes up information. It is also working on technology that can help detect the provenance of an image, Murati said. Suleyman, CEO of the company behind Pi, said another problem is that Pi and other AI chatbots aren’t designed to doubt themselves, which makes it hard to know when they’re wrong. He suggested that a possible safeguard for users would be to have responses ranked by their accuracy. “This skill of uncertainty estimation is a critical part of intelligence and actually key to making them reliable,” he said. Suleyman said he and his peers are also discussing the potential risk of AI interfering with next year’s U.S. presidential elections, and he hopes to build parameters that will prevent Pi from recommending political candidates. One of the leading risks to the development of the nation’s AI sector is the imbalance between public and private sector investment in what will soon be a technology as ubiquitous as electricity, said Fei-Fei Li, co-director of the Stanford Institute for Human-Centered Artificial Intelligence and a former vice president at Google. U.S. government investment and incentives should at least match the U.S.’s investment in space exploration decades ago with the National Aeronautics and Space Administration. “This technology is as big or even bigger than the space technology,” Li said. “We cannot just leave it to the private sector.” Li said the Food and Drug Administration, Environmental Protection Agency and other government agencies should urgently take a role in regulating AI. “It is very hard to imagine one ring that rules them all,” Li said. Roblox CEO David Baszucki said the gaming company is treading carefully when it comes to training artificial intelligence models, and isn’t harvesting anyone’s code without permission. “That’s a big societal discussion right now,” he said. The energy costs associated with powering artificial-intelligence programs have also been a concern for climate advocates. But former Meta Chief Technology Officer Mike Schroepfer, whose new firm Gigascale Capital invests in climate-focused companies, said AI will save energy in other ways. “It will be a large consumption of power, but you also have to think of the replacement costs,” he said, referring to efficiencies that AI is expected to provide. Cryptocurrency is another area in tech rife with pitfalls. Anthony Scaramucci, founder of SkyBridge Capital, said he should have been more wary of Sam Bankman-Fried, the founder of collapsed cryptocurrency exchange FTX, who is on trial in New York facing fraud and conspiracy charges. “I took the aura of all of that too seriously, and I probably should’ve been more of a skeptic,” Scaramucci said. “He committed a crime and I believe he has to go to jail for a very long time.” Executives and advocates also highlighted the risks of social media, especially for young people. Larissa May, founder and executive director, of nonprofit #HalfTheStory, said kids are spending an average of eight hours a day on their devices. “We better be looking at the place where they’re spending more time than anywhere else in their life, including sleep,” May said. Social-media companies should think about more than how much time young people are spending on social media app—they should find ways to measure whether apps are supporting or hurting them. “It’s so much bigger than just a dollar sign,” May said. Comedian and creator Elsa Majimbo said social media can be too negative. She called X, the social-media platform formerly known as Twitter, “a soft dark web” that should have a minimum age requirement of 18. Award-winning musician John Legend, who is launching his first-ever tech startup, agreed that AI has its limits. He said computer-generated music won’t replace songwriters, in part because audiences like the artists’ stories behind their music. “There’s just something that’s still so human about music, songwriting and that interaction we have with our audience,” he said. Arnold Schwarzenegger said he is aware of AI use in Hollywood and has heard a fake version of his voice. Whether or not his likeness will be used by AI in the future is a point his children will have to negotiate, the 76-year-old actor said. “I will not be around, even though I want to live forever,” he said. Even venture capitalist Khosla has tried his hand at it. When his daughter got married earlier this year, he asked ChatGPT to turn a speech he wrote into rap lyrics and then turned those lyrics into a song through an AI startup called Splash. He blared the song over speakers. “It extended my capabilities,” he said. “It meant a lot to me.” In addition to the uncertainties of AI, technology leaders are also now dealing with critical questions regarding the impact of geopolitics on the sector. Venture capitalist Khosla said that winning the race to develop advanced AI would give the U.S. an economic and political advantage over China. “I think the world’s political system—what influences Southeast Asia, Africa, Latin America—is at stake,” he said. “Western values are at stake in this technology battle, so we should do whatever we can to win this battle and beat China at it.” Khosla also warned against making the code behind advanced AI models available to the public, which some technologists have championed as a way to bolster the technology’s development. “You don’t open-source the Manhattan Project,” he said, referencing America’s clandestine efforts to build an atomic bomb during World War II. The war between Hamas and Israel, which has been a tech hub for years, was also a focal point of TechLive this year. Palmer Luckey, founder of defense technology company Anduril Industries, said U.S. corporate chiefs should be more vocal in their support for Israel. “It reflects very poorly on our billionaire class that you aren’t seeing a whole-of-country effort to become involved and to speak up about these issues. That you are seeing hedging on the condemnation of Hamas for fear of saying the wrong thing either in the court of public opinion or because it hurts their business interests,” Luckey said. Charlie Shrem, general partner of Druid Ventures, was asked about the use of cryptocurrency by Hamas to fund its attacks in Israel. He said it is “a really sad thing to see something that we were all involved in creating early on become used in these negative ways.” When it came to domestic politics, Schwarzenegger said aging leaders of the Republican and Democratic parties should step aside and make room for a new generation. The former California governor alluded to recent instances of Senate Minority Leader Mitch McConnell freezing and falling silent, and said people in that position should “start thinking about stepping aside and letting a newer generation step in and fill the vacuum.” Sarah E. Needleman, Annie Gasparro, Berber Jin, Mengqi Sun, Georgia Wells, Sarah Krouse, Heather Somerville, Tom Dotan and Deepa Seetharaman contributed to this article.

Erik Mitteregger har den 18 och 19 september genom bolaget Erik Mitteregger Förvaltning köpt 5 416 aktier i brandskyddslösningsbolaget Firefly där han är styrelseordförande och storägare.

Finansprofilen Erik Mitteregger, största ägare tillika styrelseordförande i Firfly, har på måndagen köpt aktier i bolaget för knappt 2 miljoner kronor.

Den tidigare hedgefondförvaltaren och mångårige styrelsemedlemmen i Kinnevik, Erik Mitteregger, fick fyr på aktien i sitt brandskyddsbolag Firefly i veckan. Aktien steg kraftigt efter bolagets halvårsrapport. Men det finns en sak som ligger som brandskum över den heta tillväxten: ”Rekrytering är den begränsande faktorn”, säger den mångårige storägaren och ordföranden i bolaget Erik Mitteregger till Di.

Erik Mitteregger har den 24 augusti genom bolaget Erik Mitteregger Förvaltning köpt 5 000 aktier i brandskyddslösningsbolaget Firefly där han är styrelseordförande och storägare.

Firefly, som är verksamt inom brandskyddslösningar, redovisar ett rörelseresultat för det andra kvartalet 2023 på 10,8 miljoner kronor (2,8). Rörelsemarginalen landade på 9 procent (4).

Konkurrensen om data att träna AI-systemen på hårdnar – och nu är den inte längre gratis, skriver The Economist. När efterfrågan på data ökar ser forum som Reddit och Twitter, numera X, till att ta betalt för innehållet på sina plattformar. Bolaget bakom Chat GPT, Open AI, skrev nyligen ett avtal med nyhetsbyrån AP och bildbanken Shutterstock om att få tillgång till deras respektive arkiv med texter och bilder. Samtidigt menar experter att det så snart som 2026 inte längre kommer finnas mer högkvalitativ text på internet för AI-systemen att träna på. Feeding ever-larger models is requiring makers to get creative By The Economist 8th month, 2023 Not so long ago analysts were openly wondering whether artificial intelligence (AI) would be the death of Adobe, a maker of software for creative types. New tools like dall-e 2 and Midjourney, which conjure up pictures from text, seemed set to render Adobe’s image-editing offerings redundant. As recently as April, Seeking Alpha, a financial news site, published an article headlined “Is AI the Adobe killer?” Far from it. Adobe has used its database of hundreds of millions of stock photos to build its own suite of ai tools, dubbed Firefly. Since its release in March the software has been used to create over 1bn images, says Dana Rao, an executive at the company. By avoiding mining the internet for images, as rivals did, Adobe has skirted the deepening dispute over copyright that now dogs the industry. The firm’s share price has risen by 36% since Firefly was launched. Adobe’s triumph over the doomsters illustrates a wider point about the contest for dominance in the fast-developing market for ai tools. The supersized models powering the latest wave of so-called “generative” ai rely on gargantuan amounts of data. Having already helped themselves to much of the internet—often without permission—model builders are now seeking out new data sources to sustain the feeding frenzy. Meanwhile, companies with vast troves of the stuff are weighing up how best to profit from it. A data land grab is under way. The two essential ingredients for an ai model are datasets, on which the system is trained, and processing power, through which the model detects relationships within and among those datasets. Those two ingredients are, to an extent, substitutes: a model can be improved either by ingesting more data or adding more processing power. The latter, however, is becoming difficult amid a shortage in specialist ai chips, leading model builders to be doubly focused on seeking out data. Demand for data is growing so fast that the stock of high-quality text available for training may be exhausted by 2026, reckons Epoch ai, a research outfit. The latest ai models from Google and Meta, two tech giants, are believed to have been trained on over 1trn words. By comparison, the sum total of English words on Wikipedia, an online encyclopedia, is about 4bn. It is not only the size of datasets that counts. The better the data, the better the model. Text-based models are ideally trained on long-form, well-written, factually accurate writing, notes Russell Kaplan of Scale ai, a data startup. Models that are fed this information are more likely to produce similarly high-quality output. Likewise, AI chatbots give better answers when asked to explain their working step-by-step, increasing demand for sources like textbooks that do that, too. Specialised information sets are also prized, as they allow models to be “fine-tuned” for more niche applications. Microsoft’s purchase of GitHub, a repository for software code, for $7.5bn in 2018 helped it develop a code-writing AI tool. As demand for data grows, accessing it is getting trickier, with content creators now demanding compensation for material that has been ingested into AI models. A number of copyright-infringement cases have already been brought against model builders in America. A group of authors, including Sarah Silverman, a comedian, are suing Openai, maker of Chatgpt, an AI chatbot, and Meta. A group of artists are similarly suing Stability AI, which builds text-to-image tools, and Midjourney. The upshot of all this has been a flurry of dealmaking as AI companies race to secure data sources. In July Openai inked a deal with Associated Press, a news agency, to access its archive of stories. It has also recently expanded an agreement with Shutterstock, a provider of stock photography, with whom Meta has a deal, too. On August 8th it was reported that Google was in discussions with Universal Music, a record label, to license artists’ voices to feed a songwriting AI tool. Fidelity, an asset manager, has said that it has been approached by tech firms asking for access to its financial data. Rumours swirl about AI labs approaching the BBC, Britain’s public broadcaster, for access to its archive of images and films. Another supposed target is JSTOR, a digital library of academic journals. Holders of information are taking advantage of their greater bargaining power. Reddit, a discussion forum, and Stack Overflow, a question-and-answer site popular with coders, have increased the cost of access to their data. Both websites are particularly valuable because users “upvote” preferred answers, helping models know which are most relevant. Twitter (now known as X), a social-media site, has put in place measures to limit the ability of bots to scrape the site and now charges anyone who wishes to access its data. Elon Musk, its mercurial owner, is planning to build his own ai business using the data. As a consequence, model builders are working hard to improve the quality of the inputs they already have. Many ai labs employ armies of data annotators to perform tasks such as labelling images and rating answers. Some of that work is complex; an advert for one such job seeks applicants with a master’s degree or doctorate in life sciences. But much of it is mundane, and is being outsourced to places such as Kenya where labour is cheap. AI firms are also gathering data via users’ interactions with their tools. Many of these have some form of feedback mechanism, where users indicate which outputs are useful. Firefly’s text-to-image generator allows users to pick from one of four options. Bard, Google’s chatbot, similarly proposes three answers. Users can give Chatgpt a thumbs up or thumbs down when it replies to queries. That information can be fed back as an input into the underlying model, forming what Douwe Kiela, co-founder of Contextual ai, a startup, calls the “data flywheel”. A stronger signal still of the quality of a chatbot’s answers is whether users copy the text and paste it elsewhere, he adds. Analysing such information helped Google rapidly improve its translation tool. There is, however, one source of data that remains largely untapped: the information that exists within the walls of the tech firms’ corporate customers. Many businesses possess, often unwittingly, vast amounts of useful data, from call-centre transcripts to customer spending records. Such information is especially valuable because it can be used to fine-tune models for specific business purposes, like helping call-centre workers answer customers’ queries or business analysts spot ways to boost sales. Yet making use of that rich resource is not always straightforward. Roy Singh of Bain, a consultancy, notes that most firms have historically paid little attention to the types of vast but unstructured datasets that would prove most useful for training ai tools. Often these are spread across multiple systems, buried in company servers rather than in the cloud. Unlocking that information would help companies customise ai tools to better serve their specific needs. Amazon and Microsoft, two tech giants, now offer tools to help companies better manage their unstructured datasets, as does Google. Christian Kleinerman of Snowflake, a database firm, says that business is booming as clients look to “tear down data silos”. Startups are piling in. In April Weaviate, an ai-focused database business, raised $50m at a value of $200m. Barely a week later PineCone, a rival, raised $100m at a valuation of $750m. Earlier this month Neon, another database startup, raised an additional $46m in funding. The scramble for data is only just getting started. © 2023 The Economist Newspaper Limited. All rights reserved.