Has the emergence of DeepSeek at all changed your IT department's plans for Gen AI? What do you think the future holds for OpenAI and all the other major tech players? Any advice for next steps from here?
Sort by:
I think AI, in general has been riding a growth bubble, and this is just one of the ways that those unrealistic expectations are reduced. It happens with every product and innovation. There is a Gartner Hype Cycle report for any of these technologies that shows this trend repeating itself for many years.
Therefore, we are staying the course, knowing that there are “real” benefits and business savings to be had from this technology. Just be cautious not to over invest, or over commit. Have reasonable goals and expectations. What is wrong with exceeding some conservative milestones?
No changes to our strategy.
It's really just another tool.
Being a government organization it's on the no fly list for us.
It's good to see new approaches and designs that encourage the ongoing innovation of the technology as well as increased competition that will ultimately result in a price adjust mode by existing key players in the marketplace. Hopefully it will also result in moving us closer to the commoditization of the technology.
No it hasn't, DeepSeek is just another tool - how it solves business problems remains to be seen. Which is the same view as OpenAI, Gemini, etc.
From a stock market perspective, my view is that DeepSeek probably caused a correction shares like Nvidia which were likely suffering from Artificial Inflation (the other AI).
As for next steps, treat GenAI like every other tool - don't get caught up in the hype, but ask "what problems will this solve?"
No, we haven't changed our plans. DeepSeek is good for business and competition, regardless of what they may ultimately turn out to be, positive or negative. Evaluate this model like you would any of the other models currently in play, with the understanding that China is behind its development, and whatever implications that may have within your company or personal dealings with AI and LLM's.
I think DeepSeek is innovation on how to optimize a more complex LLM’s training and weights down into a simpler less complex model. However, I think that DeepSeek’s approach and innovation does not expand the capabilities paradigm at all that we are seeing with O3, Grok, or Gemini. Given this, I don’t think it should materially change the slope or velocity of anyone’s plans in the AI space.
Also, the cost saves and other benefits they claimed they achieved is all coming under scrutiny now and being rebuffed and rebutted.
Also, I think there was a geopolitical point at play with there announcement to:
1. Call into question the effectiveness of the GPU sells restrictions.
2. Cause alarm and pause in America’s extraordinary progress in AI so they can catch up.
This was a classic move straight out of Sun Zsu’s Art of War.