Last night’s changes in AI have been seismic post the shock resignation of Sam Altman.
It is still early days and these changes will be played out. Undoubtedly, this change will impact AI roadmaps worldwide. So, how should you re-calibrate your AI road map post Sam Altman leaving OpenAI? Has anything really changed?
I was always interested in OpenAI – mainly for the reason that OpenAI was (and I believe is) – the only company with the stated goal of achieving AGI. Its too early to see the internal implications at OpenAI – but I am still interested in the wider question of AGI adoption.
The primary implication of the work from LLMs is – Is scale alone enough? This is the fundamental question on which all other questions are dependent. We see some very intriguing behaviour from LLMs based on scale (which could be evidence of emergence). But in any case, this is the primary issue on the road to AGI.
In this sense, not much has changed. OpenAI has demonstrated that this problem (of scale in AI) is worth pursuing. And many will (who have deep pockets also) – like Meta and Nvidia will pursue this question.
I will share more in future posts but,
a) There is a move to regulate the L in the LLM – I think it will not get the benefits of LLMs and will lead to competitive disadvantage to countries who do so.
b) It’s possible that due to the investments in AI, state actors may now step up. This is not a good move in my view. It could affect the competitive positioning of countries – especially the USA. My only hope is – AGI should remain out of state actors and with liberal democracies.
So, to conclude, I think the AGI genie is out of the bottle -and not much has changed – even if new leaders emerge.
However, broader questions remain. These I shall cover in following blog posts
- Leaving aside the issues with OpenAI – if a new leader emerges – who will it be?
- What is the implication for open source and AI ?
- What does it mean for OpenAI competitors?
- What does it mean for startups?
My personal belief is that AI will pursue scale – but will look at bigger / moonshot strategies in science and business – and less on the consumer strategies that need access to data.