Tech Brew Ride Home - Wed. 08/30 – OpenAI Making Bank

Episode Date: August 30, 2023

The FBI took down a huge botnet. OpenAI is apparently ahead of schedule in terms of making a ton of revenue. Why Samsung wants to apply AI to your refrigerator. Why OnStar is doubling down on AI. And we officially have the deets on the iPhone event scheduled for next month. Links: Qakbot botnet dismantled after infecting over 700,000 computers (BleepingComputer) UAE launches Arabic large language model in Gulf push into generative AI (FT) OpenAI Passes $1 Billion Revenue Pace as Big Companies Boost AI Spending (The Information) Samsung debuts its own 'AI-powered' smart recipe app (Engadget) GM is using Google’s AI chatbot to handle simple OnStar calls (The Verge) Apple Announces 'Wonderlust' Event Expected to Feature iPhone 15, Apple Watch Series 9 and More (MacRumors) Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the Tech meme right home for Wednesday, August 30th, 2023. I'm Brian McCullough today. The FBI took down a huge botnet. Open AI is apparently ahead of schedule in terms of making a ton of revenue. Why Samsung wants to apply AI to your refrigerator, why OnStar is doubling down on AI. And we officially have the deets on the iPhone event scheduled for next month. Here's what you miss today in the world of tech. On the law enforcement beat, on August 25th, the FBI apparently led an effort to dismiss.

the CakeBot botnet. I'm assuming it's Cakebot. It's QA-K-B-O-T, which ransomware gangs used as an infection vector for years. The FBI achieved this after infiltrating its network, quoting bleeping computer. Cakebot, one of the largest and longest running botnets to date, was taken down following a multinational law enforcement operation spearheaded by the FBI and known as Operation Duck Hunt. The botnet, also known as QBot and Pink SlipBot, was linked by law enforcement to at least 40 ransomware attacks against companies, health care providers, and government agencies worldwide, causing hundreds of millions of dollars in damage, according to conservative estimates. Over the past 18 months alone, losses have surpassed

$58 million. Throughout the years, CakeBot has consistently served as an initial infection vector for various ransomware gangs and their affiliates or operators, including Conti, ProLock, Egregor, Reval, Ransom X, Mega Cortex, and most recently Black Basta. The victims ranged from financial institutions on the East Coast to critical infrastructure government contractors in the Midwest to a medical device manufacturer on the West Coast. FBI director Christopher Ray said, the FBI dismantled Cakebot after it infected over 700,000 computers, over 200,000 in the United States after infiltrating parts of the botnet's infrastructure, including one of the computers used by a Cakebot admin. On Friday night, they directed the bots traffic to servers controlled

by the agency, which provided the FBI with the access needed to deploy an uninstaller to compromise devices across the globe, clearing the infection, and preventing the deployment of additional malicious payload. While victims received no notification when the uninstaller was executed to remove the malware from their systems, the FBI notified them using IP addresses and routing information collected from the victim's computers when deploying the removal tool. Furthermore, people can check if their devices were infected by submitting their email addresses on, Have I Been Poned, or the Dutch National Police websites. The FBI also worked with CISA, Shadow Server, the Microsoft Digital Crimes Unit, the National Cyber Forensics and Training Alliance, and Have I Been Poned to notify victims. The operation

was coordinated by the FBI's Los Angeles Field Office, the U.S. Attorney's Office for the Central District of California and the Criminal Division's computer crime and intellectual property sections in cooperation with Eurojust. Cakebot was the botnet of choice for some of the most infamous ransomware gangs, but we have now taken it out. This operation has also led to the seizure of almost $9 million in cryptocurrency from the cybercriminal organization, which will now be made available to victims, said U.S. Attorney Martin Estrada, end quote. There is a new major LLM in the world. Say hello to Jace. an open-source bilingual LLM for Arabic speakers, quoting the Financial Times.

An artificial intelligence group with links to Abu Dhabi's ruling family has launched what it described as the world's highest quality Arabic AI software as the United Arab Emirates pushes ahead with efforts to lead the Gulf's adoption of generative AI. The large language model known as Jase is an open-source bilingual model, available for use by the world's 400 million-plus Arabic speakers built on a trove of Arabic and English language data. The model unveiled on Wednesday is a collaboration between G-42, an AI company chaired by the UAE's National Security Advisor, Sheikh Tanun bin Zaid Al-Nahin, Abu Dhabi's Mohamed bin Zayed University of Artificial Intelligence, and Cerebras, an AI company based in California.

The launch comes as the UAE and Saudi Arabia have been buying up thousands of hyperbollah. performance Nvidia chips needed for AI software amid a global rush to secure supplies to fuel AI development. The UAE previously developed an open source large language model known as Falcon at the state-owned Technology Innovation Institute in Mazdar City, Abu Dhabi, using more than 300 Nvidia chips. Earlier this year, Cerebras signed a $100 million deal to provide nine supercomputers to G-42, one of the biggest contracts of its kind for a would-be rival to Nvidia. Jace performed better than Falcon, as well as open-source models such as Lama, when benchmarked on its accuracy in Arabic. According to its creators, it has also been designed to have a more accurate understanding of the culture and context of the region,

in contrast to most U.S.-centric models, said Professor Timothy Baldwin, acting provost of M-B-Z-U-A-I. He added that guardrails had been created to ensure that Jace, quote, does not step outside of reasonable bounds in terms of cultural and religious sensibilities, end quote. I'm sharing this story, one, because it's interesting, one of the largest LLMs yet launched anywhere in the world, but also because it brought up a bunch of questions in my mind. You know how for years you could become the Google of Russia or the Uber of Southeast Asia, say, like you could clone a successful tech product from somewhere else in the world by focusing it on a given geographic region that that product hadn't yet penetrated? So is such a thing possible here too?

can someone make a big Spanish large language model next? China's obviously working on their own versions of generative AI for their own populations and sensibilities, but also, and clearly people smarter than me must have thought about this already. How long will these LLMs literally be based on language? Language is just language, right? In a way, it's just code. It's just signifiers for stuff, for thoughts. It seems dumb that training AI on Dutch or Farsi might require a whole different model or might get different results than, say, in English or a French model. Could you build an LLM that maybe is trained on a given language, but in practice works no matter what output language you would want? Or better than that,

it could understand the underlying what, the symbology, the code behind a given language itself in order to produce the LLM without the language itself? Again, these are probably dumb questions, but it makes you think. Let's stick with AI for a bit. A source is telling the information that Open AI is on pace to generate more than $1 billion in revenue over the next 12 months, largely by selling software and computing capacity to its various products. That's up from just $28 million in revenue last year. Quote, that's far ahead of revenue projections the company previously shared with its shareholders according to a person with direct knowledge of the situation. The billion dollar

revenue figure implies that the Microsoft-backed company, which was valued on paper at $27 billion when investors bought stock from existing shareholders earlier this year, is generating more than $80 million in revenue per month. OpenAI generated just $28 million in revenue last year before it started charging for its groundbreaking chatbot chat GPT. The rapid growth in revenue suggests app developers and companies, including secretive loans like Jane Street, a Wall Street firm are increasingly finding ways to use OpenAI's conversational text technology to make money or save on costs. Microsoft, Google, and countless other businesses trying to make money from the same technology are closely watching OpenAI's growth. The company's bottom line

couldn't be learned, but it lost around $540 million last year as it developed GPT4 and chat GPT. Features involving large language models generally require servers with special chips that draw more power than the servers powering traditional software features. OpenAI CEO Sam Altman did not have a comment. The percentage of revenue OpenAI generates from chat GPT subscriptions versus selling access to GPT4 through an application programming interface also couldn't be learned. But in March of this year, OpenAI had between 1 million and 2 million chat GPT subscribers paying $20 per month, said a person with knowledge of the figure. For traders at a top trading firm, such as Jane Street, who are aiming to make profitable bets in markets,

AI language models can help by quickly parsing large amounts of data or paraphrasing lengthy memos, giving traders answers that can inform their rapid-fire decisions, according to people close to Jane Street and other similar companies. Another high-frequency trading firm Citadel has gotten close with OpenAI and other large language model developers as it looks to incorporate the technology the information reported in June. Jane Street, for instance, is a customer of OpenAI's GPT4 API, The hedge fund has reserved dedicated server capacity for the software, according to one of the people with knowledge of the company.

That implies that it paid top dollar for a faster version of OpenAI's cutting-edge model, as Altman told some developers earlier this year. Only customers who commit to spend at least $100,000 annually can get dedicated capacity according to someone present during the discussion. That capacity runs on Microsoft's cloud servers, and it isn't clear what percentage of the revenue OpenAI gets to keep from those deals. precisely how much Jane Street is spending couldn't be learned, end quote. Which points to the larger question of, how is this AI stuff actually being used in the real world? Where is the value accruing? Does it all just go to the big LLMs and their backers, open AI in this case? Are there real businesses in this AI stuff, or is it all just parlor tricks? Like we've asked many times before, you know that I have an interest in this for investing reasons,

but even if there was no fund, I'd be interested in thinking about this because of my obsession for 30 years about how the tech industry evolves and to talk about that on the show. So let's go to the other end of the spectrum for two other examples of AI being deployed. First, Samsung has launched food, a personalized AI-based food and recipe app in eight languages and 104 countries, drawing on more than 160,000 recipes from whisk, which Samsung bought in 2019. Quoting a gadget. As it promised last week, Samsung has launched food, a personalized AI-powered food and recipe app in eight languages and 104 countries around the world. It draws on the food database of WISC, an app Samsung acquired a few years back, and

resembles a version of WISC the company revealed last year. Given Samsung's large presence in kitchens with its smart fridges and other appliances, the release of a food and recipe app seems a logical step for the company. The app allows users to search for recipes around the world, save them, and make weekly eating plans. The company prepared over 160,000 recipes for launch, with that number set to increase down the road. Samsung Food can also be run on mobile phones and Samsung Family Hub smart appliances like refrigerators while allowing users to manage ingredients, shopping, etc. Users can save recipes any time and the app analyzes them, standardizes the format, and organizes them into shopping lists based on the ingredients.

It can also provide recipe recommendations based on available food items and managed by the user. It even has a personalized recipe function that uses the AI to alter recipes and create vegan or vegetarian versions, for instance. Users will even be able to create fusion recipes such as Korean versions of Italian dishes and adjust cook time or skill level of recipes. Samsung adds. The app uses AI to create recommendations for individualized daily meal plans based on dietary preferences and favorite cuisine types. Nutritional ingredient breakdowns can be viewed at any time and users can add items to shopping lists and then send them straight

to a retailer's e-commerce checkout. With connected cooking, it lets users pre-heat ovens, set timers, and transfer cook settings to supported appliances via a step-by-step guided cooking mode. Last week, Samsung said it hoped to secure a million users for the app around the world. While there are numerous recipe apps out there, meal time, paprika, yumli, etc., Samsung may have an edge with the millions of its smart appliances sold, making it a known quantity to consumers. Samsung plans to add new features like integration with Samsung Health to sync parameters like BMI and calorie consumption, while offering suggestions for diet management. The app will incorporate AI Vision Tech by 2024, allowing Samsung food to recognize food items

through the camera and provide details like nutrition information. Samsung Food is now available for download on Android and iOS, end quote. And look, this is one of the dead obvious use cases that people have been talking about from day one of this AI revolution, the slow death of the call center. GM has announced an extension of its Google Cloud Partnership to use AI, leveraging chatbots to handle simple on-star calls and freeing up operators to address more complex requests. Quoting the Verge. GM offered more details about its use of AI to coincide with Google's Cloud Next Conference.

GM had been working hand-in hand with the tech giant on a number of fronts, including Google's built-in infotainment software and OnStar. The Automaker introduced its OnStar Interactive Virtual Assistant in 2022, which utilizes Google Cloud's conversational AI technologies to provide responses to common inquiries, as well as routing and navigation assistants. And GM is already planning for future uses of AI in its vehicles. Generative AI has the potential to revolutionize the buying ownership and interaction experience inside the vehicle and beyond, enabling more opportunities to deliver new features and services.

Mike Abbott, GM's executive VP for Software and Services said in a statement, GM has extended its collaboration with Google Cloud and the development of the tech company's dialogue flow, which will allow OnStar's virtual assistant to handle more than one million customer inquiries a month. The technology is now available in the U.S. and Canada in most model year 2015 and, newer vehicles with OnStar. AI is used to handle mostly simple requests like turn-by-turn navigation, but the OnStar Virtual Assistant is trained to recognize certain words or phrases that might indicate an emergency and route the call to a train specialist instead. The shift to AI has decreased wait times and, according to GM's market research, led to mostly positive reactions.

The chatbot has also freed OnStar specialist, quote, to spend more time with customers with requests that require a human touch, GM said, end quote. And finally today, just putting this on your radar, or really your calendar, I guess, Apple has announced a September 12th event at 10 a.m. Pacific time at Apple Park entitled Wonderlust, where four iPhone 15 models, the Apple Watch Series 9, and more are expected, quoting Mac rumors. While the event is expected to be pre-recorded, Apple is inviting members of the media to campus to watch the presentation in person and to presumably have hands-on time with the new devices after the products are introduced. Alongside the new iPhone 15 models, we are expecting

the Apple Watch Series 9 and a new version of the Apple Watch Ultra. Apple plans to stream the event live on its website on YouTube and through the Apple TV app, end quote. As far as the usual T-leaf reading goes in terms of looking at the event card, quoting Ureditor on X, I might as well point out the obvious here. The colors used for this, they match up with the 15 Pro colors, not exactly subtle if you ask me, L.O.L. And quote. Also, the texture of the image in the invite suggests the rumored titanium frame. The pro models are supposed to be getting this year. And as for that tagline Wonderlust, quoting Ben Mayo, they are running out of ideas for the event slogans. Huh. Nothing for you today. Hopefully on our way to Sleeping Bear Dunes when you hear this. Talk to you

tomorrow.