Gen AI — how should we think about Open Source and the winning Microsoft role

Colin Henderson

In typical form for AI, events are moving at hyper speed however we have some clarity this cold yet sunny Monday morning.

How did we get here. My interpretation follows.

ChatGPT was released late 2022 by an unheard of company OpenAI, CEO Sam Altman
Subsequent flurry of releases on Gen AI by hundreds of entities, private and public. Refer HuggingFace.
OpenAI was held up as carrying the Open Source flag alongside the efforts of politicians in North America, Britain and Europe to regulate.
The confusion over Open Source entered hyper drive with Microsofts $13bn invesymeny in Open AI

“OpenAI was created as an open source (which is why I named it “Open” AI), non-profit company to serve as a counterweight to Google, but now it has become a closed source, maximum-profit company effectively controlled by Microsoft. Fortune Mar 17, 2023

Meta further muddied the waters releasing Llama LLM with a tiered structure based on measurement of tokens in billions, alongside related restrictions based on non-competition with Meta proprietary LLM models.

As a banker looking at the corporate structure alone, never mind the Open Source matter (more to come) this was now a playground for lawyers to discern ownership of code and what is proprietary particularly in regard to Meta. But in fact that scenario never came to pass, overtaken by a proliferation of “do good” Gen AI offerings reflecting political concerns and Anthropic handled this extremely well coming from seemingly nowhere.

But hardly nowhere

Anthropic’s CEO says why he quit his job at OpenAI to start a competitor that just received billions from Amazon and Google. Dario Amodei quit his job at OpenAI because he wanted to build a more trusted model. In 2021, he founded Anthropic with his sister Daniela and other former employees of OpenAI. Fortune Sept 26, 2023

Analysis

This is the American start up model at its best in my view. The anomalies and confusion noted above needs resolution. The EU model would require years of discussion, the British model would never get resolved and await others, but in California model it takes a weekend.

Nadella has made some brilliant tactical moves, which are somewhat outside the traditional Microsoft / Gates model and his speed has provided opportunities for Microsoft to lead in Gen AI and be first to generate revenue through association of Gen AI with their products and services.

Nadella has created a model that has left Google, Apple and Meta in the dust, for now.

Of course events will move at pace now and will look different in six or twelve months, but the steps taken by Nadella are clear and foundational:

ownership of the major Gen AI company
telent retention of experienced Gen AI engineering team to populate the new “advanced research team” noted in the FT article.
leadership snd direction of the Gen AI space

Meanwhile the OpenAI boffins, the Open Source idealists led by Ilya Sutskever, the brilliant mind behind many of OpenAI’s scientific successes is meeting business reality. I have to acknowledge his efforts and desire to maintain purity of product and embedded security for consumers.

I always appreciate that idealistic motivation and have been there myself. Then business events and decisions overtake the idealism. The lesson I have learned is that being right and being practical require adjustment and reconsideration of which hill to die on. This could be a future post.

Conclusion

Microsoft, Altman and most important the engineers and data scientists are winners here. Open Source is a topic for another day but as of now it looks dead to me.

Regulators will be confused. With no Open Source leader they are left to talk with industry leaders. To coin a phrase the Gen AI space looks more like the banking space now, and his is where Gen AI needs a Dimon type leader. Such leadership has Nadella as a nascent candidate, and his approach to date augurs well.

I personally remain commited to the concept here that we are discussing Generative AI. AGI remains elusive and awaits a “transformer” moment.

Artificial general intelligence (AGI) is a theoretical pursuit in the field of artificial intelligence (AI) research that is working to develop AI with a human level of cognition Investopedia

Thank you for reading and stay tuned. I will publish to my Medium publication Bankwatch.

tags #openai #altman #bloomberg #FT #Nadella

Bankwatch