Professional Bear Poker. "Former" 9-Time Microsoft MVP. Creator of RapidStart Apps. US citizen living on a mountain in Brazil
I don't get Marc Benioff's recent attacks on Copilot, but is he right? ?? Two things come to my mind. First, no one has yet fully cracked the code on AI for business use cases, though many are getting closer. It is clearly a continuous work in progress by all parties, including Microsoft and SFDC. While both are making gains, it is way too early for either to proclaim "Victory". Second, people who live in glass houses should avoid throwing stones. Though not unusual, it's not like Benioff lobs insults at his primary competitor every day. No, something just fires him up occasionally. It is also not out of character for him. On the other hand, Satya Nadella saying he is "Making Google Dance" is quite out of character for him. I think it just points to the stakes involved with AI. There is a palpable fear of losing the race from all participants. And even though they are all in their garages working on their cars, throwing a little oil on their competitors' track lines in advance is just a future-thinking strategy. I'm sure Benioff would welcome a public duel with Nadella over this, but that's never gonna happen.
Organizations who have relied on the lack of data discoverability for security, without proper data security controls and data classification, will need to adapt to the fact that Copilot (and many other AI tools for that matter) make your data infinitely more discoverable. CIO’s and CISO’s should be paying attention to this and looking at their data access across the organization. I also think that Benioff is acting like his product has fallen behind Microsoft, AgentForce came out about a year after Copilot. It’s not a good look ??
It feels like every one is battling for the same thing and in the end everyone will create basically the same thing. In a couple years will there really be a huge difference in anyone’s AI? Think about the early days of when Ford and GM were battling to create cars. In the end the car just became a commodity everyone needed with no big differences between the 2.
No Marc Benioff has clearly been wrong related to some of his attacks on #Copilot. He may have some valid points; however, they are lost when he makes claims that are incorrect.
Paradoxically, a company that stays a bit behind now and does not overinvest may rebound and stay afloat when people realize the emperor is naked. It doesn't matter if it's Copilot, Einstein or another o1234 model - the use cases have already reached their peak potential in normal business and they are: content summarization, content generation, activity summaries (Not for “normal” ML; for LLMs). Therefore, despite all the marketing claims (someone needs to get back the invested money), my opinion has been and is and will be the same - there will be no AGI, and LLMs will not bring any ROI. Betting everything on them will kick back. Satya should stop overinvesting just right now to keep Microsoft alive in 2-3 years, but he is all in - he will cut partner incentives, fire thousands of people and raise prices of all products, just to be first... First in what? In Copilot genereting email tones and email content, because no other serious use cases will ever benefit from any LLM. And people are already seeing this and are slowly having enough. All “spectacular” use cases are only marketing from OpenAI. The funny thing is that they must be so ashamed of the overpromise for o1 that they now jump to o3 with even bigger promises.
Don’t think anyone has significant advantage. But I would mind own busines rather than trashing competitiion. Personally, I will trust brands who focus on fair work to show their best without put down others.
The headlines are fun but ignore them. Define your business needs, evaluate tools, implement them to meet your goals and then iterate.
Founder and CEO at Delta Bravo AI | Using AI to restore leadership in Manufacturing, Water Quality
3 个月Its a slow moving luxury liner gradually changing course; the factors behind the scenes (people shortage, better data collection/quality, the realization that siloes inhibit innovation) are forcing businesses to create the foundations that will eventually align to AI business cases. They'll hit the low value ones first that align/integrate with entrenched systems of record, but the real money will be made when businesses figure out how to create AI-driven intellectual property from their processes and historic data. It's happening slower than any of the hyperscalers would like, but we will get there.