Due-diligence once again proving to be mostly cargo-culting.

I’m still learning this advantages and differences between them, would there be benefits to SFT and RAG? Or does RAG make SFT redundant?

I think generally, SFT is like giving the LLM increased intuition in specific areas. If you combine this with RAG, it should improve the performance or accuracy. Sort of like being a lawyer and knowing something is against the law by intuition, but needing the library to cite a specific case or statute as to why.

Thank you I appreciate the reply and that analogy helps make sense of this.

You can break it down by, new construction, planned renovations/improvements, and emergency repairs.

Not everyone works all three or wants to do more than one of these groups. There’s different levels of demand, pay, competition at each.

Instead of slapping on an extra battery pack, it will be an onboard llm model. Could have lifecycles just like phones.

Getting bigger (foldable) phones, without losing battery life, and running useable models in the same form-factor is a pretty big ask.

I could see a future integrated feature among cars brands doing this with their cameras. Teslas in your area on the road are already monitoring everything, then flag available street parking spots so autopilot cars looking for parking are dispatched to the nearest one. No need for drones if you already have a sufficiently large fleet of cars on the road.

Why wouldn't we just wire up all the parking spaces with sensors and then have a city wide parking app just for this? That seems far more efficient and safer over having a fleet of parking scout drones randomly flying throughout the city.

Because the car owner has an incentive to spend a couple thousand extra for a scout drone, but cities have very little incentive to spend many millions on a major infrastructure project that achieves the same. Despite that being cheaper than everyone getting a drone port in their car

It feels intuitive that cities would have the same incentive to maximize use of parking spots as they do having them in the first place. Cities want commerce to flow.

But let’s simplify it down a bit further: pretend all parking spots are for-profit. These lots would want to communicate vacancy to maximize use. Much like how motels are motivated to tell you when they have vacancy without you having to stop to find out.

The correct answer seems to be: because our economic model requires that new junk be constantly invented in order that we have something to buy. Alas.

BO definitely does layoffs, and is run just as awfully as Amazon (look who the CEO of BO is as of 2024). Doesn’t matter if you’re in an office or on the manufacturing floor, the hours and demands are terrible there. Everyone I know that has worked there echoed the same problems that Amazon had.

On the last earnings call Zuck empathized how much revenue is being allocated to salaries and RSUs. I take those types of comments during earnings calls to be signals to investors that cuts will be coming to bring expenses down by the next quarter. How many rounds between earnings is speculation, but you know at at least one big one.

A friend who works in a growing division within Meta also said last week they are re-orging again.. this time flattening out everything and making first level SDMs go back to IC. It’s always been a cut throat final-stage capitalism place to work. Nothing they do should be surprising.

My company doesn’t enforce AI usage but for those who choose to use it, every month they highlight the biggest users. It’s always non-tech people who absolutely don’t understand how LLMs work and just run a single chat for as long as possible before our system cuts them off and forces them into a new chat context.

"Can't fix stupid"