Tech Brew Ride Home - Tue. 08/08 – UK Election Hack?

Episode Date: August 8, 2023

Hackers accessed the UK Electoral Commission, so how worried should we be? Apple, Samsung and others can’t wait to buy Arm stock. OpenAI explains how its bot is crawling the web. More details on the new M3 chips Apple is testing. And Apple Music finally edges closer to the one thing Spotify still does so much better. Sponsors: Notion.com/ride Links: Millions of UK voters’ data accessible in cyber-attack, says Electoral Commission (The Guardian) Apple, Samsung to invest in Arm as it eyes September IPO (NikkeiAsia) OpenAI Launches GPTBot With Details On How To Restrict Access (SearchEngineJournal) Zoom can now train its A.I. using some customer data, according to updated terms (CNBC) Even Zoom Is Making People Return to the Office (NYTimes) Apple Tests M3 Max Chip, Setting Stage for Most Powerful MacBook Pro Yet (Bloomberg) Apple Music debuts new algorithmic Discovery Station radio (AppleInsider) 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 Tuesday, August 8th, 2023. I'm Brian McCullough today. Hackers access the UK Electoral Commission, so how worried should we be? Apple, Samsung, and others can't wait to buy

armstock. OpenAI explains how its bot is crawling the web. More details on the new M3 chips. Apple is currently testing. And Apple Music finally edges closer to the one thing. Spotify still does so much better than anyone else. Here's what you miss today in the world of tech. The UK Electoral Commission is reporting a hack by what it is calling hostile actors, which began in August of 2021, and wasn't identified until October 2022. The full scope of the hack isn't conclusively known, and, you know, I've mentioned many times the high bar I have for sharing hacking stories with you, but, you know, hacking of an electoral system definitely clears that bar, quoting the Guardian. The Electoral Commission revealed it had been hacked but was, quote, not able to know conclusively what information had been accessed and said it apologized to people whose information was accessible to the hackers. It said, however, that the largely paper-based process of elections meant it would be very hard for hackers to influence the outcome of a vote. The incident was

identified in October 2022, but attackers first accessed the commission's systems in August 2021. The commission's chief executive, Sean McNally, said, UK's democratic process is significantly dispersed, and key aspects of it remain based on paper documentation and counting. This means it would be very hard to use a cyber attack to influence the process. Nevertheless, the successful attack on the Electoral Commission highlights that organizations involved in elections remain a target and need to remain vigilant to the risks to processes around our elections, end quote. The attackers were able to access reference copies of the electoral registers held by the Commission for Research purposes and to enable permissibility checks

on political donations. The registers held at the time of the cyber attack include the names and addresses of anyone in the UK who is registered to vote between 2014 and 2022 and the names of those registered as overseas voters. The registers did not include the details of those registered anonymously. The commission's email system was also accessible during the attack. McNally said, quote, we regret that sufficient protections were not in place to prevent this cyber attack. Since identifying it, we have taken significant steps with the support of specialists to improve the security resilience and reliability of our IT systems, end quote. This is interesting. NICA Asia is reporting that Apple, Samsung,

NVIDIA, and Intel all plan to invest in ARMS' upcoming IPO as soon as shares are listed in September. SoftBank, you might remember, owns 75% of ARM right now, and the Vision Fund 25%. But clearly, for strategic reasons, all these huge players feel like they need to own a piece as well. quote, arms share flotation is slated for mid to late September. The company is sounding out investor demand and hopes that arms corporate value will rise further. The chip designer's valuation has doubled since SoftBank Group bought the company for $24 billion or $31 billion at the time in 2016. The Vision Fund will sell 10 to 15% of its arm shares on the open market.

Arm plans to welcome the big chipmakers as medium to long-term shareholders selling them stakes for a few percent each. This is meant to stabilize, the stock price at the time of the listing. Arm is a critical player in the global semiconductor industry. U.S. Chipmaker, NVIDIA previously proposed a takeover of the UK chip designer. By holding arms shares, chipmakers will hope to have sway over arms management. Arm is headquartered in Cambridge in southeastern Britain. The unit has been called the country's crown jewel, having been created through a collaboration between British industry and academia. Since its founding in 1990, it has developed integrated circuit design data, the blueprint for

semiconductors, chipmakers base their products on these blueprints. Arms power-sipping designs have given it a global market share of more than 90% for chips used in smartphones where battery life is at a premium. Arms sales reached $2.8 billion in fiscal 2022 up 70% versus fiscal 2016 when SBG acquired the startup. More than 250 billion chips using its designs have been shipped in all. SoftBank Group chairman and CEO Masayoshi Son told shareholders at his company's annual general meeting in June that he expected that number to, quote, reach one trillion, end quote. We've been talking about AI flooding the web and the world with AI content and the problems

that might mean for, well, everyone, but also what it might mean for training future LLMs. If you get that recursive training AI on AI problem, well, two stories kind of on that on finding fresh training data. Open AI has detailed GPT bot, which the company says it is currently using to crawl the web for fresh data to improve its AI models and how site admins can opt out using robots.com, quoting search engine journal. Recognizable by the following user agent token and the entire user agent string, this system scours the web for data that can enhance AI technology's accuracy, capabilities, and safety. You'll have to click through the piece to get the user agent token. Reportedly, this should filter out any paywall restricted sources,

sources that violate OpenAI's policies or sources that gather personally identifiable information. The utilization of GPTBot can potentially provide a significant boost to AI models. By allowing it to access your site, you contribute to this data pool, thereby improving the overall AI ecosystem. However, it's not a one-size-fits-all scenario. OpenAI has given web admins the power to choose whether or not to grant GPTBot access to their websites. If website owners wish to restrict GPTBot from their site, they can modify their robot's text file. Regarding the technical operations of GPTBot, any calls made to websites originate from IP address ranges documented on OpenAI's website. This detail provides added transparency and clarity to web admins about the traffic source on their sites.

OpenAI's latest news has sparked a debate on hacker news around the ethics and legality of using scraped web data to train proprietary AI systems. GPTBot identifies itself so web admins can block via robots.com, but some argue there's no benefit to allowing it, unlike search engine crawlers, that drive traffic. A significant concern is copyrighted content being used without attribution. ChatGBTGPT does not currently cite sources. There are also questions about how GPTBot handles licensed images, videos, music, and other media found on websites. If that media ends up in model training, it could constitute copyright infringement. Some experts think crawler-generated data could degrade models if AI-written content gets fed back into training. Conversely, some believe OpenAI has the right to use public web data freely, likening it to a person.

and learning from online content. However, others argue that open AI should share profits if it monetizes web data for commercial gain, end quote. And hey, if you do a Zoom, you might soon be helping them train their AI. Zoom has updated its terms to let the company train its AI models using some customer data, but does not plan to use audio, video, or chat content without consent. Quoting CNBC, the service-generated data that Zoom can now use to train its AI includes customer information on product usage, telemetry and diagnostic data and similar content or data collected by the company. It does not provide an opt-out option. This isn't an uncommon data category for companies to use for these purposes, but the new terms are a measured step towards Zoom's own AI ambitions.

Customer content such as messages, files, and documents do not seem to be included in this category. In a follow-up blog post, Zoom clarified that, quote, for AI, we do not use audio, video, or chat content for training our models without customer consent. The key phrase is, without customer consent. In June, Zoom introduced two new generative AI features, a meeting summary tool, and a tool for composing chat messages on a free trial basis for customers who can decide whether or not to use them. But when a user does enable these features, Zoom has them sign a consent form, allowing Zoom to train its AI models using their individual customer content, end quote. Speaking of Zoom, from the everybody back to the office file, Zoom, which obviously was key for

helping millions of people work from home during the pandemic, is now asking, its own employees living within 50 miles of an office to work in person at least two days per week. Or as this New York Times headline puts it, even Zoom is making people return to the office. Quote, we believe that a structured hybrid approach, meaning employees that live near an office, need to be on site two days a week to interact with their teams, is the most effective for Zoom. A company spokesperson said, we'll continue to leverage the entire Zoom platform to keep our employees and dispersed teams connected and working efficiently, end quote. during a tense meeting last week about the return to office policy held on Zoom.

Eric Yuan, the chief executive, faced a series of questions from employees who expressed frustration about the time and money they'd waste while commuting, according to an employee who was at the meeting but was not authorized to speak publicly about internal company matters. In 2020, participants in daily Zoom meetings leaped to over 300 million from 10 million the year before as it became the most downloaded free iPhone app of the year. But the company has struggled to maintain its pandemic growth. In February, amid a wave of layoffs across the tech industry, Zoom cut 15% of its staff or about 1,300 people. The company's workforce has grown more than 275% between July 2019 and October

2022. On an earnings call in May, Mr. Yuan said he was confident in the future of workplace flexibility and the benefits it had brought for his company. I think hybrid work is going to stay, he said. Zoom, like many other tech companies, is still holding on to some flexibility, requiring its employees to come in only on a part-time basis. Hybrid and remote work levels remain far above what they were pre-pandemic. As of July, nearly one-third of the country's full-time workers were in hybrid arrangements, spending some days working from home and some in an office, according to researchers at Stanford, end quote.

Developer logs are revealing that Apple is indeed testing an M3 Max with a 16-core CPU, a 40-core GPU, and 48 gigabytes of RAM, in a new MacBook Pro that Apple is expected to launch in 2024, quoting Bloomberg. The central processing unit includes 12 high-performance cores, which handle demanding tasks like video editing and for efficiency cores for less intensive applications, such as browsing the web. Compared with the top-of-the-line version of the current M2 line for laptops, the new chip has four more high-performance CPU cores and at least two additional graphics cores.

The MacBook Pro in testing also includes 48 gigabytes of memory. The company is likely testing multiple variations and core count options of which this version is one. A representative for Cupertino-California-based Apple declined to comment. The M3 chip will mark the first time Apple is shifting to a 3-nometer production process for Mac chips, which promises to provide better battery life and strong performance gains. Apple will use a similar technology for the A-17 processor launching in the iPhone 15 Pro in September. As with earlier Mac chip generations, the company is preparing a range of different M-3 models. The base M3 will have the same configuration as the M2, with eight CPU cores and up to 10 core

graphics cores, according to test logs. An M3 Pro chip, meanwhile, will start off at 12 CPU cores and 18 graphics cores. The M3 Max chip is the next level up, and an even higher end chip. The successor to the M2 Ultra hasn't shown up yet in test logs. The transition to the M3 chip will likely begin on October, starting with updates to the company's entry level max. Apple has been testing M2, 3-based iMacs, 13-inch MacBook Pro's, 13- and 15-inch MacBook Airs, and Mac Minis, all for release within the next 12 months. Updated versions of the 14- and 16-inch MacBook Pro models are likely to come in 2024 with the M3 Pro and M3 Max chips, end quote.

And finally today, speaking of Apple again, Apple Music has launched Discovery Station, akin to Spotify's algorithmic suggestions available in the Listen Now tab in the Stations for you category. Quoting Apple Insider. Finding new music can be a challenge, especially on a service like Apple Music, which hasn't caught up with Spotify's algorithmic suggestions. But a new Discovery Station may help bridge that gap. Apple doesn't add new algorithmic playlist or stations often. So the new personal radio option arrived as a surprise on Monday. The Discovery Station appeared for users sometime early, Monday, August 7th, without any official announcement or pop-up. Users browsing the Apple Music

app can see the new option under the Listen Now tab in the category station. for you. Apple Insider found this new playlist thanks to a small change to the personal station that's been around for years. The artwork for the personal station changed from a basic red play button to an animated red arrow pointing to the right. This curious change led us to believe something else might have been updated, and that's when we saw the Discovery Station in Stations for you. Its artwork is the inverse of personal stations with a blue color scheme and an inverted arrow animating inwards. We can only guess how it functions based on the personal station without any documentation or explanation of this new station. The Apple Music

Personal Station appears after asking Siri to play music that I like and is an algorithmically built station based on play history, likes, and other factors in the Apple Music algorithm. As its name implies, the Discovery Station appears to be an algorithmically generated station built to play music the user hasn't heard but may like. So it functions like the new music mix playlist, but can play continuously. This kind of continuous listening experience for music discovery built on an algorithm has been missing from Apple Music so far. Users could run stations based on genre, artists, or songs, but those didn't account for the user's taste. It seems that Siri isn't yet aware of this new radio station.

At least, we haven't discovered what command is required to get it to play, end quote. Reminder that if folks want more information about the Ride Home AI Fund, Chris and I will be happy to talk to you, jump on the phone, jump on a Zoom, email you a deck, give you more details, etc. Hit me up at Brian at Ridehomefund.com. to chat. Talk to you tomorrow.