"First, all of the weaknesses of LLMs are being addressed by exponential increases in compute"
Are we sure about that? Because I have not seen evidence on this. I kinda subscribe to LeCun's argument that LLMs cannot be "fixed" with more compute. But I am open to any counter-argument and evidence to the contrary.
"Second, the number of people who need to wield AI effectively for demand to skyrocket is decreasing."
I am not even sure what this means. What does "wield AI effectively" mean? What does the whole sentence mean?
"Third, the economic returns from using agents aren’t just impactful on the bottom line, but the top line as well."
What evidence do we have on AI helping companies generate more revenue organically because of its capabilities, not because of the demand for components? Google and Facebook, likely. The same article even said that consumers did not care about agents when it came to buying Apple products. Yes, we have heard Apple is behind in AI for months and guess what, more people are warmed up to the idea that the company might be smart all along.
"what has Apple done to inspire faith that they can do a better job?"
Or maybe because consumers don't care about agents or AI just because business executives, analysts or influencers like the author keep talking about agents and AI? If AI can make a product better, great! But should Apple focus on delivering the best products and experience possible whether or not it involves AI? Or should it deliver AI at all costs?
Anthropic found a product market fit, a clientele that finds value in paying for an LLM, whether it's real efficiency or just an excuse to cut costs. OpenAI must continue to ship models. That's how it keeps the music going. But a lot of consumers don't want to pay for something that doesn't justify the monthly cost that is just going to get more expensive.
"First, all of the weaknesses of LLMs are being addressed by exponential increases in compute"
Are we sure about that? Because I have not seen evidence on this. I kinda subscribe to LeCun's argument that LLMs cannot be "fixed" with more compute. But I am open to any counter-argument and evidence to the contrary.
"Second, the number of people who need to wield AI effectively for demand to skyrocket is decreasing."
I am not even sure what this means. What does "wield AI effectively" mean? What does the whole sentence mean?
"Third, the economic returns from using agents aren’t just impactful on the bottom line, but the top line as well."
What evidence do we have on AI helping companies generate more revenue organically because of its capabilities, not because of the demand for components? Google and Facebook, likely. The same article even said that consumers did not care about agents when it came to buying Apple products. Yes, we have heard Apple is behind in AI for months and guess what, more people are warmed up to the idea that the company might be smart all along.
"what has Apple done to inspire faith that they can do a better job?"
Or maybe because consumers don't care about agents or AI just because business executives, analysts or influencers like the author keep talking about agents and AI? If AI can make a product better, great! But should Apple focus on delivering the best products and experience possible whether or not it involves AI? Or should it deliver AI at all costs?
Anthropic found a product market fit, a clientele that finds value in paying for an LLM, whether it's real efficiency or just an excuse to cut costs. OpenAI must continue to ship models. That's how it keeps the music going. But a lot of consumers don't want to pay for something that doesn't justify the monthly cost that is just going to get more expensive.