AI Breakthroughs: Claude Opus 4, OpenAI's $6.5B Deal, and Atera's IT Autopilot for 40% Workloads

[00:00:02] It's Friday, May 23, 2025, and I'm Dave Soule. Four things to know today. Anthropics Claude Opus 4 challenges rivals with autonomous reasoning and developer transparency, OpenAI grabs headlines with a $6.5B hardware play and real substance in new MCP support, Ateras IT Autopilot claims to automate 40% of IT workloads, and Verizon's $20B merger approval comes with a politically charged twist.

[00:00:29] Plus the big ideas. Google bets on personalization while critics ask if AI should know everything about you, and Microsoft's Aurora aims to forecast the world. This is the Business of Tech. Well, it's been quite the week of AI announcements. And not to be outdone, Anthropik has unveiled its latest artificial intelligence models Claude Opus 4 and Claude Sonnet 4, designed to enhance coding capabilities and problem-solving skills.

[00:00:57] The company claims that Claude Opus 4 is the most powerful model to date, capable of running autonomously on long tasks for several hours, and outperforming competitors like Google's Gemini and OpenAI's model encoding tasks. According to Anthropik's internal benchmarks, Claude Opus 4 is 65% less likely to take shortcuts compared to its predecessor, demonstrating improved accuracy and efficiency. The new models also feature thinking summaries to clarify the reasoning processes,

[00:01:25] and an extended thinking mode that enhances performance. OpenAI has announced its acquisition of I.O., a hardware company founded by former Apple design chief Johnny I., in a deal valued at nearly $6.5 billion. This acquisition will bring approximately 55 hardware engineers and software developers into OpenAI, as they aim to enhance the design and development of new technology. Johnny I.O. will not join OpenAI directly,

[00:01:53] but his design firm, LoveFrom, will continue to operate independently while overseeing design of all of OpenAI's products. The first devices resulting from this collaboration are expected to launch in 2026, and they will not aim to replace existing technology, but will represent a new type of product. OpenAI has announced significant updates to its Responses API, introducing support for remote model context protocol servers, native image generation capabilities, and enhanced enterprise features.

[00:02:23] These updates, which went live on May 21st, aimed to streamline the development of intelligent applications for businesses. The Responses API, launched in March 2025, has already processed trillions of tokens and supports diverse applications, ranging from market research to software development. With the new enhancements, developers can now integrate external tools like Stripe and Shopify easily, build applications that generate and edit images in real-time, and execute complex data analysis tasks.

[00:02:53] OpenAI has also introduced features such as improved file search capabilities and reasoning summaries to enhance transparency and efficiency for their enterprise users. Although one is more a no-AI feature, Signal has announced a new update for its Windows application that will prevent Microsoft's recall feature from capturing screenshots of secure chats addressing privacy concerns. The update enables screen security by default,

[00:03:18] a feature similar to digital rights management that blocks screenshotting on services like Netflix. Signal's developer, Joshua Lund, emphasized the need for operating system vendors like Microsoft to provide developers with the tools necessary to protect sensitive information from artificial intelligence systems. Despite previous delays in launching recall, the feature raises significant security and privacy issues, as it does not currently allow app developers to opt out of having sensitive content archived by its AI.

[00:03:48] Lund pointed out that while Microsoft filters certain activities by default, users must know how to adjust settings to protect their information effectively. Why do we care? Anthropix Claude Opus 4 may be their most powerful model yet, but model iteration is table stakes now. Let's expect it even. Improved benchmarks and thinking summaries are nice for devs, but the real value for service providers lies in how these tools integrate into workflows and remain secure. That's where other announcements this week become more important.

[00:04:18] Let's be real. OpenAI's acquisition of Johnny Ive's Startup.io is a $6.5 billion valuation headline on a deal that didn't even include Ive himself. The hardware engineers are real, but this smells more like a marketing move than an execution milestone. Their new kind of device planned for 2026 could be interesting, but we've seen this move before. Magic leap, the human AI pin, I could go on.

[00:04:44] The real under-the-radar game changer is OpenAI's extension of MCP support in its responses API, now allowing remote MCP server integration. MCP was introduced as a vendor-neutral interface for AI agents to plug into operating systems and software with controlled access to tools, files, and contacts, and now OpenAI is enabling it for production use. Note, I've talked about it several times this week, as Microsoft and Google also announced support.

[00:05:12] Signal's move to block Microsoft Recall from screenshotting secure content is a key signal in the growing pushback from developers. While Microsoft intends Recall to enable AI-driven search across user activity, it is currently opt-in, but it's poorly documented and introduces serious privacy risk, especially in regulated environments. AI's value isn't in smarter models, but in safer, more transparent integrations into business environments. Today's episode is supported by Huntress.

[00:05:40] Most cybersecurity solutions are built from massive enterprises with big budgets, not Huntress. They're the fully-managed cybersecurity platform built for all businesses, not just the 1%. Huntress purposely builds security solutions like EDR, ITDR, SIM, and security awareness training to equip their team of elite threat hunters to handle the heavy lifting of security for you.

[00:06:03] When threat actors strike, Huntress's 24x7 Global Sock shuts them down before they're even on anyone else's radar. But they do more than just chase alerts. They lead the charge in industry research and knowledge, bringing expert protection and peace of mind. That's why users on G2 rate their EDR number one for growing businesses. To see how their expert threat hunting team gets the job done, visit Huntress.com slash MSB Radio.

[00:06:34] Atera has launched IT Autopilot, designed to automate the resolution of Tier 1 IT tickets without human oversight. According to Atera's recent survey, 85% of IT professionals believe that a 4-day work week would be achievable if Tier 1 tickets were eliminated from their responsibilities. Atera's IT Autopilot can handle up to 40% of IT workloads, achieving response times as low as 0.1 seconds and average resolution times of just 15 minutes.

[00:07:02] Gil Peckleman, co-founder and CEO of Atera, emphasized that this technology not only automates tasks, but also restores work-life balance for technicians by allowing them to focus on more impactful projects. Kaseya has partnered with Pulseway to enhance their offerings for IT professionals, providing integrated solutions and access to the Kaseya ecosystem. This collaboration aims to leverage both companies' strengths to deliver advanced tools for managing and securing IT environments.

[00:07:28] With over 13,000 businesses, including notable names like Best Buy and Dell, utilizing Pulseway's mobile-first remote monitoring and management solutions, the integration is expected to enhance product capabilities and support. Why do we care? Atera's new IT Autopilot claims it can automate up to 40% of IT workloads, including those resolutions of Tier 1 tickets without human oversight with 15-minute average resolution times. The messaging targets technician burnout and a reduced work week's fantasy,

[00:07:56] which is clever marketing, but it's the underlying automation performance that matters. The automation that creates new classes of errors or misses customer nuance can backfire fast. Kaseya's partnership with Pulseway appears to be a strategic ecosystem expansion,

[00:08:22] integrating Kaseya's broader product line, particularly its security offerings, into Pulseway's RMM, which enjoys significant traction among mobile-first IT teams and S&P administrators. Unless you're a Pulseway user, this isn't transformative. However, it highlights a growing platform tension in the market, open ecosystems versus consolidated mega-platforms. Service providers should continue to evaluate integration control against vendor lock-in, particularly regarding cybersecurity dependencies.

[00:08:52] Two acquisitions to talk about. Proofpoint has made a significant move in cybersecurity by acquiring Nuclei, a company specializing in compliance archiving and artificial intelligence-driven data enrichment for workplace communications. This acquisition comes shortly after Proofpoint's announcement of its acquisition of Hornet Security, a UK-based cybersecurity provider for over a billion dollars. With the integration of Nuclei, Proofpoint aims to enhance its digital communications governance offerings,

[00:09:19] enabling businesses to capture, retain and analyze communications across various collaboration platforms, including Microsoft Teams and Slack. The Federal Communications Commission has approved Verizon's $20 billion merger with Frontier Communications, marking a significant consolidation in the Tagalog communications industry. Verizon will pay $9.6 billion in cash and assume $10 billion in debt to acquire Frontier, aiming to enhance its fiber network across the U.S.

[00:09:47] This merger comes with a commitment from Verizon to discontinue all diversity, equity and inclusion initiatives, aligning with the current administration's focus on dismantling such practices in both public and private sectors. The FCC's approval is expected to facilitate substantial infrastructure development, with plans to deploy fiber to over 1 million American homes annually across 25 states. Why do we care? Well, Proofpoint's move to acquire Nuclei following its billion-dollar pickup Hornet

[00:10:16] represents a familiar pattern in enterprise cybersecurity, stacking complementary capabilities around compliance, communications and AI enrichment. But let's be real, this is a safe incremental play. No market shift here. Just continued consolidation of adjacent capabilities by a major player. The FCC's approval of Verizon's $20 billion acquisition of Frontier marks one of the largest telecom consolidations in years. On paper, it's about fiber expansion. The attached requirement to abandon all of DEI initiatives

[00:10:45] is a sharp and troubling deviation from precedent. Telecom M&A historically under-delivers AT&T and DirecTV, Sprint and D-Mobile. Consolidation often brings price hikes, slower innovation and worse support. This deal comes with a political mandate. FCC approval was conditioned on Verizon agreeing to dismantle diversity, equity and inclusion programs. That's unprecedented. The risk?

[00:11:11] A federal agency dictating corporate HR policy as a term of merger approval, undermining not only operational autonomy, but it sets a dangerous regulatory precedent. Let's end the week with some big ideas. The Verge offers perspective on Gemini's significant AI advantage. Google is leveraging its extensive data resources to enhance its artificial intelligence capabilities, particularly through its Gemini model.

[00:11:38] The company's new personal context allows Gemini to access a user's search history, emails and files from Google Drive to provide tailored personalized responses. This follows Google's earlier initiative that lets users opt into a personalized version of Gemini, designed to deliver uniquely insightful responses. According to Google CEO Sundar Pachai, the ability to analyze previous communications will enable the AI to generate replies that mimic the user's tone and style,

[00:12:07] potentially improving interpersonal interactions. This approach contrasts sharply with competitors like OpenAI, whose chat GPT starts with no prior knowledge of the user, requiring it to gather context over multiple interactions. With personal context, Google aims to position Gemini as a more intuitive AI assistant from the outset, capable of delivering relevant information without needing previous engagement. And from the New York Times, artificial intelligence is set to transform weather forecasting

[00:12:36] with the introduction of Microsoft's Aurora model, which promises faster and more accurate predictions. Unlike traditional forecasting methods that rely on complex mathematical equations and take hours to produce results, Aurora can generate 10-day forecasts quickly and efficiently, running alongside existing models at one of Europe's largest weather centers. Developed with data from physics-based models, Aurora's versatility allows it to predict not just weather patterns,

[00:13:04] but also other Earth's system events, such as air pollution and wave heights. Paris Pythagarchus, a professor at the University of Pennsylvania and a key developer of the model, emphasizes its potential to enhance forecasting systems significantly. However, experts caution that while AI models can speed up the forecasting process, they still require careful collaboration and human verification before widespread use.

[00:13:31] Jay McBain highlighting how generative AI is transforming how technology vendors engage with their channel partner ecosystems, but many organizations are not yet prepared to fully leverage its potential. According to Canalys, over 90% of partner programs remain at a maturity level that hinders their ability to implement advanced AI solutions effectively. As the global channel continues to grow, now encompassing millions of partners, there is a marked shift in the roles of channel professionals.

[00:14:00] New AI tools are being introduced to automate administrative tasks, thereby improving response times and collaboration. However, experts emphasize that foundational automation must be established before organizations can successfully adopt generative AI technologies, indicating a critical need for investment in basic data management and automation processes. Jay will be on the live show on Wednesday. You should join us and ask a question. Finally, some opinions on ChatTubt.

[00:14:27] Simon Willison echoed my recent feelings with some frustrations around the new memory features. The model now aggregates and updates a detailed summary of user conversations ongoing. Willison highlights that the ability to recall past conversations can affect the outcomes of prompts in unexpected ways, complicating research and creative tasks. Critics argue that this extensive data accumulation creates an overwhelming profile that can unintentionally influence future interactions,

[00:14:55] particularly for users who engage in diverse topics. And what do you do with all that large context in GPT 4.1? Futurepedia details a test conducted with a 161-page real estate market report, where GPT 4.1 was able to summarize key takeaways and answers direct questions about the content. According to the article, workers utilizing generative artificial intelligence tools like GPT 4.1

[00:15:21] save an average of 5.4% of their work hours each week, which translates to over two hours in a standard 40-hour workweek. Now why do we care? Your questions to ponder this long holiday weekend were off on the news on Monday for the U.S. holiday. If the best models are all good enough, what matters more? How personalized they are or how private they remain? Should we trust AI with real-world decisions? If so, what validation and oversight do we need?

[00:15:51] Is your organization ready for AI or still catching up to 2015-level automation? Should your AI know everything you've ever done with it? And as the model's converging capabilities, the battleground becomes who has the right data access, who delivers the better outcomes, and who manages risk and privacy responsibly. When chaos hits, will your system survive?

[00:16:16] Comet Backup's full image backups protect your entire system, files, apps, settings, and even the OS, in one powerful automated backup. Comet is easy to deploy and bandwidth efficient. Full system backups for Windows, for Linux, for servers, for endpoints, all in one centralized platform. Whether you get hit with ransomware, hardware failure, or accidental deletion, Comet restores your system to physical, VM, or the cloud. Short on time?

[00:16:43] Restore individual files or folders to access critical data within moments. Try Comet Backup today. Reliable, secure, and made to scale. Visit cometbackup.com to protect what matters most. Get $100 free credit when you sign up with the promo code MSPRADIO. Comet Backup. Backup. Fast. Restore. Easy. Thanks for listening.

[00:17:07] Today is National Taffy Day, National Cooler Day, National Road Trip Day, and National Title Track Day. Those are some wacky ones. It's also the day before a long U.S. holiday weekend, and I'll be back with you on Tuesday. The Business of Tech is written and produced by me, Dave Sobel, under ethics guidelines posted at businessof.tech. If you've enjoyed the show, make sure you've subscribed or followed on your favorite platform. It's free and helps directly.

[00:17:37] Give us a review, too. If you want to support the show, visit patreon.com slash MSP Radio, and you'll get access to content early. Or buy our Why Do We Care merch at businessof.tech. Have a question you want answered? We take listener questions, send them in, ideally as a voice memo or video, to question at mspradio.com. I answer listener questions live on our Wednesday live show on YouTube and LinkedIn.

[00:18:05] If you've got a comment or a thought on a story, put it in the comments if you're on YouTube, or reach out on LinkedIn if you're listening to the podcast. And if you want to advertise on the show, visit mspradio.com slash engage. Once again, thanks for listening, and I will talk to you again on our next episode. Part of the MSP Radio Network. We'll be right back. Bye.