> small differences were important enough to change the world unpredictable ways
I completely agree with this. I’m also in my 50s but I lean techno-optimist. Whenever there’s a new hype cycle, I’m looking for those small differences that might overcome the hurdles we’ve seen in the past. For example, the PicturePhone was demoed in 1964. Now we’re all on Zoom constantly. I’ve lived through many AI hype cycles. The current LLM craze is significantly different from past exuberance. LLMs aren’t perfect but they are remarkable when they work.
I don’t know where you live but I get the battery swapped while I’m standing there. It’s insane that you’d rather spend $800 than be without your phone for a few hours.
I live in the US. I've had friends without their phone for days due to a bad swap. I wasn't really talking about myself (I have a spare phone at all times I can simply switch to in an emergency) - but the average person doesn't have spares laying around.
It works until it doesn't.
It's also rarely spending "$1,000" - it's usually a $200-400 price difference after trade-in depending on how often you upgrade. Apple typically gives extremely aggressive pricing on trade-ins, or at least has over the past few years. I sent my wife in for a battery swap and it ended up being $150 more to simply swap phones and upgrade 3 generations to reset the obsolescence clock. No-brainer at that point.
Professional managers (aka not the founder, climbed up the corporate ladder the old fashioned way): Tim Cook (apple), Satya Nadella (msft), Sundar Pichai (Google), Arvind Krishna (ibm), Shantanu Narayen (adobe), Dara Khosrowshahi (Uber), Lisa Su (amd) and on and on. Just scan the Fortune 500 for tech companies and it’s nearly all professional managers. The YC audience could have debunked this Founder Mode nonsense in just 10 minutes on Wikipedia, as I did.
Most professional managers are BS artists. Perhaps Chesky hired of bunch of those. The conclusion is not to run everything yourself, but maybe hire better managers.
The immersive videos are amazing. Stick a VR camera on the best seat at a concert, sports event, wedding, etc. People will pay for Taylor Swift or the Super Bowl. The tech is very close, the Quest is a good price, VR cameras are ~$5k. The only thing left is the content creators.
I worked on online learning for a bit. Turns out people are willing to pay for the inconvenience of in-person learning, even flying to another location. It's the only way most people can focus on a topic. Otherwise, work, kids, life interrupts and they can't stay on track. Replit's 100 Days of Python says only 0.4% of those who complete day 1 finish day 100.
Good info! I use an HPC with SLURM. 40k GPUs shared by hundreds of users. It works well enough. I don’t know how the market for cloud-based clusters works. Why didn’t OP use AWS or Google for on-demand training? Is it just down to cost?
If you do, in fact, need H100s, they can be very hard to get. Even the smaller flavors of A100 you sometimes request, wait days for, and then 1 node might show up during a weekend. And for the reasons described in the article and the fact that large training jobs can be network-limited, nicer networks can be a big deal.
I completely agree with this. I’m also in my 50s but I lean techno-optimist. Whenever there’s a new hype cycle, I’m looking for those small differences that might overcome the hurdles we’ve seen in the past. For example, the PicturePhone was demoed in 1964. Now we’re all on Zoom constantly. I’ve lived through many AI hype cycles. The current LLM craze is significantly different from past exuberance. LLMs aren’t perfect but they are remarkable when they work.
reply