In the last two weeks I read multiple article on the end of "hype" or "cooling down" of the growth cycle of GenAI. McKinsey released a detailed report on why the hype is over and the bubble can burst anytime. I believe most of it may be because of NVDA, AMD and Super Micro stock prices trending down, for the right reasons, and is not a reflection of how the tech is evolving. If anything, the growth is still on the upward segment of the hockey stick.
Few reasons why I believe the Tech is only growing faster than any other time.
- The most significant release is the new OpenAI o1 series (including o1-preview and o1-mini), designed to tackle complex tasks in fields like coding, science, and math. These models spend more time thinking before responding, which allows them to handle harder problems more effectively than previous versions.
- Llama models are approaching 350 million downloads to date (more than 10x the downloads compared to this time last year), and they were downloaded more than 20 million times in the July month alone. Monthly usage (token volume) of Llama grew 10x from January to July 2024.
- Hugging Face Hub platform has over a Million models, 200k datasets, and 300k demos in which people can easily collaborate in their ML workflows. It is growing at the rate of 3% a week.
- AI Startup companies are on track to match or exceed the roughly $21 billion they raised in 2023. I do acknowledge half of them are just selling a lottery ticket (pitch deck/term sheet) and may not even actually be creating a product.
- NVIDIA is doing well but it is the utilities that are struggling. BlackRock, Global Infrastructure Partners, Microsoft and MGX launch new AI partnership to invest in data centers and supporting power infrastructure.
- Companies are rapidly embracing the Gen AI era. Solutions like Microsoft Co-Pilot and Salesforce's AgentForce are accelerating adoption by simplifying workflows and enhancing productivity. A standout example is #BORN, a Tech Mahindra company, which leveraged Gen AI to create a full-scale production-quality advertisement—automating everything from models to product creatives. t
Critics point to a cooling off in media coverage and businesses reassessing the technology’s true potential as a maturing phase but in reality it is their attempt to slowdown the world for them to understand. This train is moving too fast to stop.
Partnering with the most innovative AI and RPA platforms to optimize back office processes, automate manual tasks, improve customer service, save money, and grow profits.
3 周Hi Minoji - fact: there is still a hype around new models and applications that pushes stock further up. Especially the big seven. fact 2:because of the lack of return of investment, the stocks are way overvalued. So this means that even though we all would like it to continue, new cash to pay for the bills might dry up (openAI 5B US loss per year). This in turn means that the stock valuation will become unstable, and a push or even a nudge can take the whole house of cards down. I witnessed the dot com burst. I see the same patterns. However, where this is different is that there is a possibility that we move from hypecycle to hypecycle instead of staying with just one hypecycle. What will happen is that when a hype will decline and start to move down into the trough, a next hype picks up and creates another peak, whilst the other hype declines. So you will basically jump from one peak to the other. And that is I think the strategy that they are employing. They (big seven) are releasing new stuff lightning fast. They should cause otherwise their house of cards will tumble down
SVP | P&L Owner | Guiding Organizations Through Growth, Innovation, and Strategic Transformation
3 周In the AI prospects world split between optimists and pessimists, I consider myself in the former camp. It is however important that more and more efforts start happening now on the application side of AI with clear focus on credible ROI path generation. The technology side of AI has seen a lot of focus and devotion but without the bandwagon now moving to the real world value identification and creation zone, the adoption will not follow the promise this technology offers.