<< Previous Edition: Is Tribal Knowledge an Antidote
It is often said that if we fail to remember history, we are doomed to repeat it. This saying takes an interesting turn in the field of software engineering, where history is bound to repeat itself, but in a more evolved form. Moreover, the recurrence is not dependent on whether you remember it or not, but those who do remember stand to benefit the most from it.
Software Engineering History in an Infinite Loop
A particularly fascinating facet of software engineering history is the selective incorporation of the best ideas and practices from diverse sources into a powerful "cocktail" of innovation. This suggests that not every player gets what they wish for. Indeed, it could be posited that each player is granted two wishes, but only one of these wishes will be fulfilled. Which wishes are these, and which ones will come to fruition?
Meta's Dual Desires
a) To set the benchmarks for the future of software engineering using Large Language Models (LLMs).
b) To pioneer open-source innovation in this field.
Meta is poised to achieve its goal of establishing standards akin to how Sun Microsystems paved the way with Java, yet without the associated challenges, given Meta's position outside the traditional technology vendor landscape.
OpenAI's Ambitions
a) Become the dominant player in the LLM market.
b) Establish itself as the de facto standard in the industry.
OpenAI is well-positioned to maintain market leadership but may face challenges in becoming the sole standard in a market eager to flex open weights.
Anthropic and Mistral's Diverse Desires
a) To present open-weight models featuring parameters accessible to the public.
b) To supply closed-weight models, preserving the exclusivity of their parameters.
Despite initial inclinations—Anthropic leaning towards closed-weight models and Mistral favoring open-weights—my perspective is that both will ultimately achieve success by embracing open-weight standards as established by Meta (in near fu true).
Cloud Providers' Dual Hopes
a) Innovate and advance LLMs.
b) Operationalize LLMs for enterprise use cases.
While cloud providers like AWS, Google Cloud, and Microsoft Azure may make some model innovations, their strength will lie in simplifying the deployment and scalability of LLMs.
Conclusion
As the generative AI landscape evolves, various players in the field are striving to shape the future of innovation according to their unique visions and aspirations. Each key player, from tech giants like Meta and OpenAI to emerging companies like Anthropic and Mistral, has distinct wishes for the direction of the industry. However, the path of progress is rarely a straight line that satisfies all desires equally.
>> Next Edition: No Winner in Gen AI will get their Full Wish
