Any issues with the standard sealing it deep in abandoned mine with many layers of security before a breach? Thankfully its not actually that much waste space wise, but logistically have to move it around/store it temporarily and thats where stuff goes wrong
The actual capital that was invested was spent optimizing how we put apps into production, and building technology to take advantage of parallelization/load balancing/scaling etc. Whereas with LLMs we need a whole new ecosystem of hardware and deployment because its a completely different use
What im trying to say is that for most standard reasons you need cloud computing, the capital investment is just not nearly as extensive/important as the technology of deployment. AWS is basically going to spend the same amount on AI data centres as it did from 2011 to 2022
i mean we are just infinitely better at most things with compute. I can spin up a searchable index for TBs worth of data in minutes and it can be incredibly optimized. CPU/process efficiency just isnt a bottleneck for a lot of tasks. ML data centres are much more about expanding to new tasks
schwarzkopf was another weird one
The brief stretch where NZ went YIMBY had 0% rent growth for 5 years. Certainly feels like we are getting the evidence for the most basic assumptions over time
The dirty secret is that their business facing parts are where the product management is
god this would kick so much ass
immortality
You can kind of get it. Like something about him is quintessentially canadian and therefore kinda relatable to the northeast. Unfortunately it also includes the very bad parts