This will keep going around the table, next it might be a Chinese company that demos 98% of the capability at 1/4 the price. The objective of being at the cutting edge of LLM performance seems like more of a marketing advantage in the game of sucking in more capital for a moatless technology.
> Which makes me think they are getting the strategy exactly backwards
The strategy is to take an admittedly cool technology and spin a dramatic story around it to raise capital, while providing a rationale for workforce reductions. Remember that investment chases stories, not actual results (whether financial or societal).
When enough capital is there, it will be considered "too big to fail". Maybe it's already there.
Which makes me think they are getting the strategy exactly backwards. My problem is usually not something that would be solved by the AI being better but instead by it being more integrated into my life.