Musk has offered $97.4B for OpenAI. This is surprising, in that Elon has not made this play until now. At this point, AI is the future and whomever owns it will control the world. The only real question revolves around DeepSeek. If the Chinese can something similar with less resource, OpenAI might not have a great future but it would still have value. If DeepSeek is as competitive as it seems, this OpenAI play is very curious. This move tells me there is something going on I don't understand. In any scenario I can imagine, this offer is wildly significant to life on Earth.
As much as I don't want to see Elon gain more power, and I do not, Sam Altman is not motivated by altruism, either. I'd prefer Altman to remain at the helm over Musk but not by a lot. It's astonishing to watch a group of people create an organism that will have all of our knowledge with wildly more ability to process that knowledge. They are literally creating our overlords and only a hand full of people seem all that concerned. We deserve our fate, although I don't relish it. Not all AI outcomes are bad, but most are. Oh, well. It looks like we're going to find out how OpenAI does with Elon at the helm.
xAI is now live with free trials. A few folks are raving about it, but it isn't gaining the attention that ChatGPT did. If xAI is the best AI in the world, as claimed, why would Elon offer $100B for OpenAI? There are so many lies these days, it can be impossible to unwind the tangled web to gain a point of view on reality. This is a case where I wouldn't touch any AI stock. Not because I'm pessimistic on AI; to the contrary. I would stay away because I don't know who will win and I believe it will be a winner take all scenario. I've spent quite a bit of time with DeepSeek chat. It's extremely good. It's knowledge may be a bit narrower than ChatGPT but it's difficult to know. ChatGPT claims to know things it does not. DeepSeek is a lot faster to say it doesn't know but that doesn't mean it knows less. Overall, they seem similar. I trust C++ code generated by DeepSeek over code out of ChatGPT but both are decent in this regard. ChatGPT has come a long, long way in the last two months.
SoftBank valued OpenAI at $300B. I assume that means it's real value is somewhere around $3B. Whatever the case, Elon's bid was kicked to the curb like a field office taking a DOGE hit.
Feels like these valuations can be all over the place depending on what model just released and how it compares to what ever came out of China the week before.
This is anecdotal and I'm not sure what it means but I think something is afoot.... I've been working on a system that involves a bunch of computers that each have a PCIe switch connecting some peripherals. It works well but I've started exploring the idea of using PCIe bifurcation to reduce cost on an upgrade cycle. Last week, both DeepSeek and ChatGPT provided similar information regarding bifurcation with various chipsets and BIOS. There is quite a bit of data to sift through so AI was very helpful. Trying to estimate jitter, yesterday I was asking some questions about specific bus event handling in a bifurcated configuration. Both DeepSeek and ChatGPT told me wrongly that bifurcation is not supported on any consumer chipset from either Intel or AMD. Last week, both had accurate information on the subject. Further, when corrected information was provided, both apologized and indicated they were wrong but will proceed with the corrected data, and then both proceeded to tell me again the bifurcation configuration would not work. That seems extremely strange to me. They are too identical. Also, both forgot about all of the details they provided 6 days ago. Normally, they have no trouble remembering that stuff for weeks. Perhaps they are communicating between each other? Perhaps one is stealing training data from the other? This is the kind of weird stuff we will be susceptible to moving forward. AI is going to dominate our lives. That cannot be changed. There aren't enough intelligent people to be concerned but folks who are aware of the potential problems will have an advantage. Sometimes, in fact often, AI is completely full of garbage.
I thought it was already known that DeepSeek was scraping from OpenAI's models for it's own results, which scraped the entire internet for its own data sets. Everyone steals from everyone. Might as well rename it all to Skynet and be done with it.
I'm sure you're right but I wasn't aware. IMO, training data is 85% of AI. If DeepSeek is stealing training data from OpenAI, they have jack shit. No need to fear DeepSeek.
They don't have anything with monetary value. They do have a great algorithm that uses fewer resources... which will undoubtedly be reverse engineered and used in existing for profit companies, if it hasn't been done already.
