Are AI Models Getting Commoditised?
Raghav Gupta
Founder & CEO at Futurense | Co-Founder at 1% Club | Democratising Opportunities for Bharat ??
AI commoditisation is something I’ve seen many people in the industry talking about lately, and it really reflects where things are headed. For AI companies, this calls for a strategic shift in how they approach AI.
Hi folks! ?? It’s been a busy few weeks, but now that I’ve got some time, let’s dive into today’s topic!
Satya Nadella recently pointed out that AI is becoming more of a commodity. The value in AI is shifting from the models themselves to how they are used. With AI models becoming more accessible and similar in performance, the real impact now lies in creating customized solutions that bring tangible benefits.?
Nandan Nilekani shares a similar view-?
“Ultimately, the models will become more commoditized, and the value will switch to the application layer and the whole stack.”
So, the question isn't about "Who has the best model?" but "How to use these models effectively to create real value?"
All the big names in AI are pushing this movement forward, and the real value now lies in how businesses apply AI to achieve results. Let’s explore what that means in today’s edition!
First, What’s Driving AI Commoditization?
I think there are several key factors at play:
How Are Big Players Responding?
Here’s how some of the major players are reacting to this shift:
What Does This Mean for Companies Building AI Solutions?
This trend calls for a strategic shift in how companies approach AI models for two key reasons:
领英推荐
So, for companies, it’s crucial to recognize these trends and focus on smart application and integration rather than just having the latest model.?
What's your take on this shift in AI? Share your thoughts in the comments below.
Until then, stay excited and stay curious!?
Signing Off, Raghav.?
Other Reads You Might Like
Check out some of the latest posts I’ve shared recently:
Enjoyed reading this newsletter? I would love it if you stay tuned for more! Subscribe now to keep up with all the latest updates and insights.
Thanks for the simple, relatable and democratized article, Raghav ! AI commoditisation will happen (if not already) - not at all layers though, like we saw in softwarae world too, when SDN came in - NFV was an offshoot of it, yet didn't eat up the entire software stack ... Similarly the layer in that Gen AI stack that's getting the taste of commoditisation is "model" - and that's where the "open source" and "democratization" words will continue to be associated with ... From AI solution (or customer backwards perspective) - what'll matter is - what's the cost of this democratization ... And cost is again offshoot of what's the fine tuning you are looking for, which is dependent on - how soon you are looking to launch / solve a particular problem from solution perspective .... So there's a cost that's associated with this democratization of Gen AI models, it's only the question of who bears the cost and what's the transactional give-up in a scenario .... Example - you may love Gemini or Llama series - and want to use it as part of "solution" stack - even just for tokenization, if not pipeline, but there's a massive cost associated to that ...