Tech Brew Ride Home - Thu. 04/24 – The AI Coding Wars Are Upon Us

Episode Date: April 24, 2025

More fallout from those EU fines yesterday. A whole slew of self driving car news. The AI Coding Wars have officially begun. The back to the office wars continue. Is Chrome worth $50 billion? And let me tell you about the AI app that wants to help you cheat at everything. Links: Apple and Meta Are First to Be Hit by E.U. Digital Competition Law (NYTimes) Uber, Volkswagen pair up to launch robotaxi service in US with self-driving, electric microbuses (TechCrunch) Windsurf slashes prices as competition with Cursor heats up (TechCrunch) Google Chrome Worth ‘Upwards of $50 Billion,’ Browser Rival Says (Bloomberg) Google forcing some remote workers to come back 3 days a week or lose their jobs (CNBC) Columbia student suspended over interview cheating tool raises $5.3M to ‘cheat on everything’ (TechCrunch) 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 Thursday, April 24th, 2025. I'm Brian McCullough today. More fallout from those EU fines yesterday. A whole slew of self-driving car news. The AI coding wars have officially begun. The Back to the Office Wars continue is Chrome worth $50 billion. And let me tell

you about the AI app that wants to help you cheat at everything. Here's what you missed today in the world of tech. Well, a bit of a reprieve. The European Commission says that Meta's Facebook Marketplace should no longer be designated under the DMA, citing fewer than 10,000 business users in 2024, but the fallout from the fines yesterday continues. Quoting the Times, the Trump administration lashed out at the ruling on Wednesday. This novel form of economic extortion will not be tolerated by the United States, said Brian Hughes, a spokesman for the National Security Council. Extra territorial regulations that specifically target and undermine American companies, stifle innovation and enable censorship, will be recognized.

as barriers to trade and a direct threat to free civil society, end quote. A White House memo from February said officials would consider retaliation if the European Union targeted American tech companies under the Digital Markets Act or the Digital Services Act, a law focused on curbing illicit online content and disinformation. Meta said it was likely to appeal the ruling, calling it an attack on American companies akin to imposing steep tariffs on their services. The European Commission is attempting to handicap successful American businesses while allowing Chinese and European companies to operate under different standards, Joel Kaplan, Meta's chief

global affairs officer said in a statement, this isn't just about a fine. The commission forcing us to change our business model effectively imposes a multi-billion dollar tariff on meta while requiring us to offer an inferior service, end quote. Apple said it would appeal the decision and accuse the commission of forcing it to make product changes that amounted to giving its technology away. Last year, the company was hit with a $2 billion EU fine for using the App Store to undercut competition and music streaming. We have spent hundreds of thousands of engineering hours and made dozens of changes to comply with this law, none of which our users have asked for, Apple said in a statement. Despite countless meetings, the commission continues to move the goalposts

every step of the way, end quote. Flurry of self-driving car news, Volkswagen and Uber have unveiled an ambitious plan to launch a commercial robotaxy service using ID buzz vehicles in U.S. cities over the next decade starting in LA in 2026, quoting TechCrunch. Initially, the service won't be driverless. The fleet of autonomous vehicles will have human safety operators behind the wheel before they go driverless in 2027, a VW spokesperson told TechCrunch that gives Volkswagen ADMT, the Autonomous Vehicle Subsidiary of Volkswagen of America up to two years to navigate the regulatory landscape in California and gain the permits required to test its autonomous vehicles and eventually operate a commercial service. Volkswagen ADMT will begin testing in Los Angeles later this year once it receives its initial testing permit from the California Department of Motor Vehicles.

The agency regulates autonomous vehicle testing and deployment in the state, and the California Public Utilities Commission handles permitting for the commercial ride-hailing component of Robotaxi services. Despite the considerable hurdles ahead, the partnership is a notable step for Volkswagen ADMT. The subsidiary publicly launched in July 2023 with an autonomous vehicle test program in Austin and a fleet of 10. and all-electric ID buzz vehicles equipped with partner Mobilize technology. Its parent Volkswagen Group, along with Ford, had hitched their autonomous vehicle ambitions to start up Argo until the two automakers pulled financial support and gobbled up its remains. Volkswagen then turned to Mobile Eye to source autonomous vehicle technology, and that relationship has deepened recently. ADMT, Volkswagen's U.S.-based effort, launched about nine months

after Argo shut down. Volkswagen in 2023 said it wasn't interested in building a dedicated ride hailing service. Still, it did appear to see a business in selling its self-driving ID buzz vans and fleet management software to other companies. Details of its partnership with Uber suggests that plan is intact. Volkswagen is not just a car manufacturer. We are shaping the future of mobility and our collaboration with Uber accelerates that vision. Christian Sanger, CEO of Volkswagen Autonomous Mobility said in a statement, what really sets us apart is our ability to combine the best of both worlds, high-volume manufacturing expertise with cutting-edge technology and a deep understanding of urban mobility needs, end quote. This is also Uber's latest AV partnership.

The ride-hailing giant has spent the past several years locking up deals with more than 14 autonomous vehicle firms across ride-hailing delivery and trucking. Uber recently launched a robotaxie service with Waymo in Austin and is about to do the same in Atlanta, end quote. Meanwhile, Tesla says it has started testing its autonomous ride-hailing service with employees in Austin and the Bay Area and has completed 15,000 miles of driving, quoting a different tech crunch piece. FSD supervised ride hailing service is live for an early set of employees in Austin and San Francisco Bay Area. The company posted Wednesday on X. FSD stands for full self-driving, which is Tesla's advanced driver assistance system available to Tesla owners via subscription

that can perform some automated driving tasks. The system, which requires the driver to keep their hands on the wheel, is not yet capable of autonomously driving. Thousands of Tesla owners already drive themselves around with supervised FSD. Tesla's announcement Wednesday centers on the addition of a Robotaxy app that will theoretically be used by non-Tesla owners to hail a vehicle in the fleet. Tesla plans to launch a Robotaxy Service in Austin in June, which the company reiterated Tuesday during its first quarter earnings call. The automaker didn't share many other details on the call, like when it expects to start charging for rides, the most color CEO Elon Musk provided, was to say that he expects to roll out 10 to 20 vehicles on day one of service in

Austin. And while Tesla made a splash last year after it debuted its cybercab, concept, a futuristic-looking robotaxy built without a steering wheel or pedals. The company is poised to begin operations with its existing portfolio of vehicles. The promotional video Tesla shared Wednesday shows a Model 3 sedan that has been fitted with a screen in the back for passengers that displays information such as estimated time of arrival, climate and music controls, and a button for an emergency stop. A disclaimer at the bottom of the video reads, Safety driver is present to supervise and only intervene as necessary. FSD supervised does not make the vehicle autonomous, end quote.

All right, we have a full-on war going on in the AI coding space. Quoting TechCrunch again, AI coding assistant startup WindSurf cut its prices across the board, it announced on Monday, touting massive savings for its users as competition with its rival cursor intensifies. WindSurf said it's getting rid of its complex system of flow action credits, which charge developers for actions its AI did in the background. It's also cutting prices for its team plans to $30 per user per month, down from $35, while making its enterprise plans much cheaper per the announcement. Winsurf product marketer Rob Howe proclaimed on X that Winsurf now has by far the best and most affordable pricing structure of all AI coding tools on the market, crediting this to Winsurf optimizing its GPU usage.

How criticized confusing competitor plans priced at $20 a month in an apparent dig at Cursor's individual monthly plan, which starts at $20, compared to windsurf's 15. The pricing overhaul comes as windurf is reportedly being considered for an acquisition by OpenAI for $3 billion. Cursor's creator AnySphere is in talks to raise at a $10 billion valuation. As TechCrunch previously reported, WinSurf is the smaller of the two coding assistant startups generating about $100 million in ARR compared to cursors $300 million. OpenAI originally wanted to buy cursor, but it's growing so quickly that it's not in the market to be sold. Although Winsurf hasn't confirmed the OpenAI acquisition reports, it has recently stepped up its public collaborations with OpenAI. For example, WinSurf's CEO Varun Mohan appeared earlier this month in OpenAI's launch video for its latest API model family,

and as part of the pricing change announcement, WinSurf is lavishing its users with another week of free and unlimited usage of OpenAI's latest GPT4.1 and 04 mini models. The big question is whether Cursor ends up cutting its own prices in response to Winsurf's revamp, that might risk a price war, making it harder for both startups to scale up profitably, end quote. More drips and drabs from the Google remedy trial. Duck, Duck, Go, CEO Gabriel Weinberg testified that Chrome is worth upwards of $50 billion if it went on the market based on back-of-the-envelope math. I have to say I'm not sure about that math, but, quoting, Weinberg, who said the estimate was back of the envelope math, based on Chrome's user base, testified as part of a three-week hearing in the DOJ's case against Alphabet's Google to determine

how to remedy the company's monopoly in internet search. Judge Amit Meta ruled Google illegally monopolized the market last year and is contemplating a package of changes proposed by antitrust enforcers. The DOJ and a group of U.S. states have argued that Google should be forced to sell off its popular Chrome web browser. The $50 billion figure given by Weinberg is higher than the roughly $20 billion value estimated by Bloomberg intelligence analysts Mandip Singh in November. Such a lofty price tag might dissuade potential buyers from making offers if Google were ordered to spin off Chrome, end quote. And here's some more interesting math. The information says OpenAI has told investors it expects revenue of $125 billion by 2029 and revenue of $174 billion by 2030,

with sales from agents and other products exceeding the those from ChatGPT by that point. Quote, the projections which would propel the 10-year-old startup's sales toward the level of Nvidia or meta-platforms today reflect rapid revenue gains from agents or AI software that can take actions on behalf of customers as well as other new products. These include those tied to free user monetization,

likely meaning money made from OpenAI's non-paying users. OpenAI ended last year at $3.7 billion in revenue nearly quadrupling sales the year prior. The San Francisco startup has been serving more than 500 million active users per week, up from 300 million in December. What the new products are is unclear. CEO Sam Altman, in an interview with Stretekery newsletter writer Ben Thompson, recently discussed the possibility of charging affiliate fees or a cut of a sale initiated from a user's search through OpenAI software, such as ChatGPT or its agents, though he said he

is reluctant to sell traditional ads on the chatbot. OpenAI chief financial officer Sarah Fryer, meanwhile, has discussed the possibility of selling ads, but told the Financial Times the company has no active plans to pursue advertising. The forecasts provide insight into why investors and a deal led by SoftBank agreed to invest $40 billion in new capital at a $260 billion valuation, 73% higher than its valuation last fall. Spokespeople for SoftBank and OpenAI declined to comment. The projections indicate executives expect the new sources of revenue could help offset rising costs. The company, which estimates it will burn $46 billion in cash over the next four years, thanks to the costs of training and running its models and other expenses such as salaries,

expects to turn cash flow positive in 2029, generating close to $12 billion in cash that year. The company also expects growth of inference costs, the cost of running AI products such as chat GPT and underlying models to moderate over the next half decade. Those costs will triple this year to about $6 billion and rise to nearly $47 billion in 2030. Still, the annual growth rate will fall to about 30 percent then. This slower growth and inference costs combined with soaring revenue is a boon to OpenAI's margins. OpenAI anticipates gross profit as a percentage of revenue rising to nearly 70% in 2020-29 from 40% last year, which was far lower than the average gross margin of 74% for cloud software stocks tracked by Meritech Capital, end quote.

The Back to Work wars continue apace. CNBC has seen internal documents that suggest several Google units have notified remote workers that their jobs will be in jeopardy if they don't show up at the office at least three days a week. Quote, Courtney Menzini, a Google spokesperson said the decisions around remote worker return demands are based on individual teams and not a company-wide policy. As we've said before, in-person collaboration is an important part of how we innovate and solve complex problems, Mencini said in a statement to CNBC. To support this, some teams have asked remote employees that live near an office to return to in-person work three days a week. According to one recent notice, employees in Google technical services

were told that they're required to switch to a hybrid office schedule or take a voluntary exit package. Remote employees in the unit are being offered a one-time paid relocation expense to move within 50 miles of an office. Remote employees in human resources are what Google calls people operations who live within 50 miles of an office must choose to work in person on a hybrid basis by this month or their role will be eliminated, according to an internal memo. said they have to return by June. Staffers in that unit who are approved for remote work and live more than 50 miles away from an office can keep their current arrangements, but will have to go hybrid if they want new roles at the company. Google previously offered a voluntary exit program

to U.S.-based full-time employees in people operations starting in March, according to a memo sent by HR chief Fiona Sikoni in February. That came after the company said in January that it would be offering voluntary exit packages to full-time employees in the U.S. in the platforms and devices group, which includes Android, Chrome, and products like Fitbit and Nest. The unit has made cuts to nearly two dozen teams as of this month. While internal correspondence indicated that remote work was a factor in the layoffs, Mancini said, it was not a main consideration for the changes, end quote. Finally today, something to catch you up on a Columbia student was suspended over a tool to cheat on job interviews. And this led them.

to raise a $5.3 million seed round for a startup called Cluley, which offers an AI tool to, quote, cheat on everything, quoting TechCrunch. The startup was born after Lee posted in a viral X thread that he was suspended by Columbia University after he and his co-founder developed a tool to cheat on job interviews for software engineers. That tool originally called InterviewCoder is now part of their San Francisco-based startup, Cluley. It offers its users the chance to cheat on things like exams, sales calls, and job interviews thanks to a hidden in-browser window that can't be viewed by the interviewer or test giver. Cluelly has published a manifesto comparing itself to innovations like the calculator and spell check, which were originally derided as cheating.

Cluelly also publishes a slickly produced but polarizing launch video of Lee using a hidden AI assistant to unsuccessfully lie to a woman about his age and even his knowledge of art on a date at a fancy restaurant. While some praised the video for grabbing people's attention, others derided it as reminiscent of the dystopian sci-fi television show Black Mirror. Lee, who is Cluley's CEO, told TechCrunch, the AI cheating tool surpassed $3 million in ARR earlier this month. The startup's other co-founder is another 21-year-old former Columbia student, Neil Schumugan, who is Cluley's C-O. Shedmugan was also embroiled in disciplinary proceedings at Columbia over the AI tool. Both co-founders have dropped out of Columbia.

The university's student newspaper reported last week, Columbia declined to comment. citing student privacy laws. Kluoli began as a tool for developers to cheat on knowledge of LEAT code, a platform for coding questions that some in software engineering circles, including Kluly's founders, of course, consider outdated and a waste of time. Lee says he was able to snag an internship with Amazon using the AI cheating tool. Amazon declined to comment on Lee's particular case to TechCrunch, but said its job candidates must acknowledge they won't use unauthorized tools during the interview process, end quote.

Nothing more for you today. Talk to you tomorrow. Thank you.