以下、OpenAIのDevDayに関するネタをネット検索させて、記事化したもの。

OpenAI's DevDay has been a topic of considerable discussion, garnering attention across the technology sector. As an experienced software enthusiast with a penchant for the evolution of computing and AI, I find this event particularly intriguing. This gathering, showcasing the latest advancements and future directions of AI technology, represents a convergence of some of the brightest minds and most innovative ideas in the field.

The significance of DevDay extends beyond a simple product announcement or a technical conference. It embodies the rapid progression and transformative potential of AI, particularly in how it is reshaping industries, influencing consumer behaviors, and challenging our traditional understanding of software development. This event is not just about presenting new tools or features; it is a window into the future of AI, offering insights into the practical applications and ethical considerations of these powerful technologies.

DevDay is a testament to the extraordinary journey from rudimentary computing systems to sophisticated AI models like ChatGPT. It highlights the leaps in machine learning, natural language processing, and neural network capabilities, which are no longer confined to academic papers but are actively being integrated into products and services affecting millions.

Furthermore, as someone who has witnessed the evolution of the software industry from its early days, the approach OpenAI is taking with ChatGPT and other AI technologies signifies a pivotal moment. It's not just about the technical prowess of these models but also about the broader implications for software development practices, the role of AI in decision-making processes, and the ethical considerations that come with such advanced technologies.

Therefore, analyzing DevDay from this nostalgic yet forward-looking perspective provides an opportunity to appreciate the profound impact AI is having on our world, while also contemplating the responsibilities that come with such power. It's a blend of admiration for technological advancement and a cautious approach towards the potential challenges and dilemmas these innovations might bring.

Originally, OpenAI, the creator of ChatGPT, was a research organization formed by Silicon Valley all-stars. It began in 2015, led by top-tier technologists like Greg Brockman, the first CTO of the FinTech sensation Stripe, and Ilya Sutskever, a disciple of deep learning authority Professor Hinton and a researcher at Google Brain, focusing on deep learning as the core of AI research and development.

In 2019, with Sam Altman, initially a representative and director of YCombinator, assuming the CEO position, the scale of research and development expanded rapidly, particularly with Microsoft in infrastructure collaboration. Amidst various research and web service releases, the early 2023 launch of the conversational AI service ChatGPT became a massive hit, marking its major debut after eight years.

The innovation of ChatGPT as a web service can be explained by two aspects of speed.

First is the speed of user acquisition. The service surpassed 100 million MAUs in just three months after its launch, a breathtaking pace not limited to consumer adoption. Within eight months of API release, 460 of the Fortune 500 companies had already developed applications using the API, marking a phase where the remaining 40 companies were being ridiculed as shameful laggards. Surpassing 90% penetration in the Fortune 500 in less than a year is nothing short of phenomenal. Consumer adoption is also impressive, with weekly active users exceeding 100 million and showing no signs of slowing down.

The second aspect is the speed of development. Apart from improvements in model performance, the pace of developing APIs and peripheral functions has been astonishing. In late March, the Plugins announced crushed those attempting to secretly capitalize on the APIs released at the beginning of the month. This time, OpenAI released Actions enabling external API integration, Assistants for persistent context and control flow in conversations, and GPTs (plural) for adding external teacher data within minutes, instantly obliterating startups that venture capitalists had invested in without much due diligence due to the FUD surrounding generative AI. To summarize Sam Altman's keynote in a tongue-in-cheek way, "To tell you the truth, shallow services reliant on GPT's API are no longer viable. Sorry for the sudden announcement. The platform for GPT developers was completed a few days ago. That was the sign of the end. Soon, anyone will be able to create services using generative AI. Be cautious when issuing term sheets."

Investors, the jesters on the innovation stage, knowingly indulge in easy pattern recognition, and the author of this newsletter is no exception. Observing OpenAI's tumultuous year, I struggled to categorize it: Is it a reiteration of Google, AWS, or Windows? However, this DevDay has finally provided an answer, albeit a straightforward and frightening one: OpenAI is akin to Google, AWS, and Windows combined. I will explain the historical parallels with these three companies.

The common ground with Google lies in the successful productization of foundational technology. The PageRank algorithm conceived by Larry Page and Sergey Brin at Stanford University became the core of Google Search, setting the foundation for future developments.

ChatGPT's technology is based on the Transformer concept, initially researched mainly by Google. The seminal paper in this field is considered to be "Attention Is All You Need" from 2017, but it took five years for this to culminate in the product ChatGPT.

The most significant difference from Google is the speed of proliferation. While it took Google several years to become the de facto standard search engine, ChatGPT became a globally used service in an instant, leaving major IT companies, including Google, in its wake.

In terms of the commonalities with AWS, ChatGPT initially started as an API. In 2020, Microsoft announced a licensing deal for GPT-3 technology, indicating a platform strategy focused on packaging and providing element technologies. Even after the successful launch of ChatGPT, the focus remains on the API.

Recalling the history of Amazon Web Services, which has become the Don Quixote of cloud services, its first product was the S3 object storage service. In the late 2000s, Google's research sparked the Big Data revolution.

Jeff Bezos, the founder of Amazon, foresaw the imminent need for economical data storage and processing, leading to the release of S3, a platform modeled on internal infrastructure for external developers. Interestingly, EC2 was a later development.

The essence of S3 and EC2's innovation lies in their availability as (REST) APIs, allowing for hardware procurement, previously requiring human intervention, to be accessible with a single command from a laptop, greatly simplifying data processing and web service expansion.

In OpenAI's case, providing an API has created an environment where GPT can be used easily without the high technical expertise and vast time required for model tuning and teacher data discovery and preprocessing. This ease of use is reminiscent of AWS's early days.

However, unlike AWS, which required a deep understanding of server technology and certain programming skills, ChatGPT allows for the creation of AI-powered services without a deep understanding of generative AI technologies.

Another common point with AWS is how both platforms have swayed the interests of entrepreneurs and investors. Many over-40 entrepreneurs have experienced anxiety when AWS's new services closely resembled their own or their investments.

AWS once had a market presence so strong that it could crush competing services with a single press release, often attacking from the bottom of the abstraction pyramid, leaving companies reliant on AWS with no recourse when AWS introduced a similar service.

The reactions to OpenAI's recent DevDay evoke memories of AWS in the 2010s. Positively speaking, OpenAI's rapid platform development speed has allowed startups to make quick strategic shifts and decisions.

This speed is the most significant difference between AWS and OpenAI. While AWS often leads with press releases, OpenAI continually announces immediately usable services. With OpenAI moving so swiftly as the leading entity in generative AI, followers are left without recourse.

Lastly, the common ground with Windows is significant. Windows held a monopolistic position as the interface between consumers and computers until the rise of iOS. IBM, failing to foresee the advent of the PC era, relinquished exclusive rights to MS-DOS in 1980.

This allowed Microsoft to sell the operating system to various PC manufacturers, propelling it into a massive IT enterprise. Dominating the OS market with Windows, its successor to MS-DOS, Microsoft attempted to enforce the use of its subpar browser, Internet Explorer, leading to a historic antitrust lawsuit.

OpenAI is not as monopolistic as Windows was in the 90s, but its overwhelming mindshare in both consumer and business contexts is strikingly similar. Given this mindshare in both consumer and business sectors, it makes sense for OpenAI to focus on developers, termed "toD" initiatives.

Naming its new feature showcase "DevDay" or "Developer's Day" is no coincidence for OpenAI. The most critical customers for OpenAI now are not the nearly fully penetrated Fortune 500 companies, nor the continuously increasing end-users, but the developers who conceive applications that further draw in businesses and consumers.

Another company that has historically centered its go-to-market strategy around developers is Microsoft. Readers over forty might remember Steve Ballmer, the former Microsoft CEO, sweating profusely while repeatedly shouting "Developers."

Like iOS today, the value of Windows as a platform lay in the numerous third-party software complementing its proprietary software, enabled by .NET and other Windows APIs. However, the rise of web browsers and the proliferation of smartphones as the new PCs marked the end of the Windows era.

Still, consciously or unconsciously, they continue to inspire subsequent companies, and OpenAI and Sam Altman are no exceptions.

Another commonality is the lukewarm distance towards open source. This stance reflects a strategic choice that aligns with both Microsoft's historical approach and the evolving business model of AI development. Microsoft, for a long time, was known for its guarded approach towards open source, prioritizing proprietary software. However, in recent years, there has been a notable shift in Microsoft's strategy, embracing open source more openly as the tech landscape evolved. This change mirrors the broader industry trend where open source is increasingly seen as a valuable asset rather than a threat to proprietary business models.

Similarly, OpenAI initially embraced a more open approach with its earlier GPT models, releasing them widely to the public. But with GPT-3 and its successors, OpenAI adopted a more controlled release strategy. This shift could be interpreted as a response to the growing realization of the commercial potential and the risks associated with powerful AI models. By controlling access to the latest versions of GPT, OpenAI can more effectively manage ethical concerns, misuse risks, and commercial interests.

This careful approach towards open source and wider distribution reflects a maturation in the AI field, where the balance between innovation, open collaboration, and commercial viability is continually evolving. Companies like OpenAI are navigating this landscape by choosing when and how to share their technologies while maintaining a competitive edge and ensuring responsible use.

Furthermore, this strategy is indicative of a broader trend in the tech industry, where companies are grappling with how to harness the collaborative benefits of open source while protecting their intellectual property and commercial interests. The challenge lies in finding a balance that fosters innovation and community engagement without compromising on business goals or ethical responsibilities.

In conclusion, OpenAI's current stance on open source, mirroring aspects of Microsoft's journey, underscores a complex interplay between open innovation, commercial strategy, and the ethical management of technology. As AI continues to advance, the strategies of companies like OpenAI in balancing these elements will likely set precedents for the future of technology development and deployment.