Time for some big ideas.
I’ll start with OpenAI – they are facing scrutiny over its chaotic internal operations as highlighted by former engineer Calvin French-Owen. In a recent blog post, French-Owen pointed out that the company’s rapid growth from just over 1,000 employees to more than 3,000 in a year has led to significant organizational challenges, including a lack of centralized planning and oversight. He noted that many teams at OpenAI tend to operate independently, often leading to overlapping projects and duplicated efforts. French-Owen described the work environment as one driven by a strong bias for action, where individual teams frequently make decisions without broader coordination. This structure has resulted in communication breakdowns and inefficiencies, reminiscent of a government research lab rather than a cohesive tech company. The exclusive reliance on Microsoft Azure for operations and the predominant use of Slack for communication further complicate these issues, raising questions about security and effective cross-team collaboration.
I thought this piece was useful for those considering the business implications of AI. 404 Media discusses the media industry’s misguided pivot to artificial intelligence, highlighting that this approach is not a viable business strategy. Despite generative AI being a significant threat to journalism, media executives continue to view it as a promising opportunity, often at the expense of their workforce. Statistics reveal a stark contrast in traffic referrals: Google has directed approximately 3 million visitors to media websites since their inception, while ChatGPT has sent only 1,600. Executives at companies like Business Insider and Hearst Newspapers have emphasized the need to embrace AI, with Business Insider’s CEO stating a “huge opportunity for companies who harness AI first.” However, the article argues that relying on AI will not solve the industry’s fundamental challenges, as quality journalism must prioritize human connection and trust over automated output.
Alex Komoroske highlights the limitations of the “same-origin paradigm,” a security measure from the 1990s that has inadvertently allowed major technology companies to consolidate power by trapping user data within their ecosystems. This paradigm prevents seamless integration between applications, making it difficult to transfer data across platforms and ultimately amplifying the dominance of large companies. Komoroske, who has over a decade of experience at firms like Stripe and Google, argues that despite the promise of artificial intelligence to democratize software development, the existing framework may actually exacerbate data silos rather than eliminate them. He points to advancements in technology, such as secure computing enclaves, that could enable a new approach where data policies are attached directly to the data itself, allowing for more flexible and secure interactions. The potential for significant change is now within reach, as these innovations could help redefine how software and data coexist in a more user-centric manner.
Why do we care?
If OpenAI—arguably the most influential AI company—can’t coordinate its own teams, how confident are you in their ability to safely manage your clients’ data and workflows?
Are your clients chasing “AI” because of real customer needs—or because it looks good to hype driven owners, investors and boards?
As browsers and cloud services become the new “OS,” is your MSP prepared to manage security and governance across these silos?
What core technology choices were made long ago that we can question now to create new opportunities?