Featured in this episode of Tech News of the Week
So one thing about online AI services like ChatGPT or DallE is that the data you put into them becomes someone else’s property- in those cases OpenAI, but insert whatever service you want and the result is the same.
What companies are starting to do is run private instances, and companies like Nvidia are all too happy to help them do it- for a price. To run private instances on the latest and greatest GPU-laden systems in the Nvidia DGX Cloud, companies can expect a starting price of $37,000/instance/month. STARTING PRICE.
To be fair, these are the bleedingest of the edge instances, hosted by Nvidia themselves, so it makes sense that they’d be crushingly expensive. AWS and Azure have their own options but if you want to get something close to enterprise grade you’re still looking at a healthy 5 digit monthly bill.
For companies like Samsung, who you’ll remember had a bit of an oopsie when their engineers gave OpenAI proprietary code, that’s a drop in the bucket. After all, can you truly put a price on peace of mind? Or, I guess to rephrase… can you truly put a price on protecting intellectual property? I guess you can, and it’s at least $37k a month.