pmart123

There's a lot of irony right now regarding the cost of these things too (although I know the cost curve will drop over time). I know developers that are burning $1,000/day on tokens for Claude Code or VC's using the $200/month ChatGPT pricing plan who are then talking about Vibe coding TurboTax away. TurboTax for most people is $50 to $100 a year. We are still a far way off even from a cost justification standpoint let alone a reliability standpoint of relying on a vibe coded solution for filing your taxes.

I remember when the hype first started around it, it was unusably slow, and produced poor results. Granted, I haven't tried it lately to see if latency improved, but the hype versus product state at the time, really turned me off from the product.

That's partially why it's hard. People are automatically unhappy if their flight is delayed, their bags take time to arrive, they have to gate check a bag, sit near a baby during the flight, etc. It would be very hard to run perfectly with the weather being such a large input factor. Even food tastes worse up 35k feet due to the lack of moisture and air pressure. Think about how unhappy people would be with Apple if every time it rained or snowed, the phone couldn't function due to physics?

It's a much harder business to run than a software company.

I'm no expert in this, but I think one common misconception among regular skiers is that a ski tune is "free" rather than each ski only gets so many full tunes as it wears down the base and edges. Obviously, if you get a core shot or something like that, it becomes more necessary. I always understood waxing skis as more protective though?

There are still many instances where this partnership is the case though. Search and rescue, avalanche rescue, herding, security and law enforcement, disability assistance, bomb/drug detection, hunting, etc. A lot of the companion dogs are useless for these roles, but many popular breeds could perform these tasks if given the training. On the west coast, people generally seem much more active with their dogs imo.

I hear your points, but can't git, CI tools, IDEs like DataGrip all be used to test and deploy the database code too versus letting some db admin create functions or procedures with no version control? Also, with things like RDS or readonly replicas, couldn't more analytical queries be done directly in the database versus busing data around?

The api interface is simple, but the change would impact the code underneath. Since these are branch and bound algorithms, it would really depend on how often the worst runtime complexity case occurred. If it only happened in 2% of use cases, it might not make a huge difference for example.

Always thought great JIT that's compatible with Cython and major projects with C extensions like numpy, scipy would be a more worthwhile effort. A lot of the data intensive tasks Python can be run in parallel processes easily so doesn't seem like a major benefit to removing the GIL?

That's most of investing though. Investors won't all agree the operating cash flow, asset base, organic sales growth if there's been M&A, etc. let alone with a P/E, P/FCF, EV/EBITDA ratio makes a company cheap or expensive. Yet, this doesn't make it political in nature though.

I don't believe the outputs are completely deterministic?

Completely agree on his take. Generative AI perhaps might be good at certain, specific tasks in terms of pulling or summarizing information for an analyst, but it won't help predict markets. It'll be much better at helping aid in software where the problem is more deterministic.

A lot of the ai plays seems like small vertical business opportunities that fill a particular niche that could be quickly profitable for a small team, but aren't really scalable or could quickly become obsolete.

I don't see it replacing H&M look books as it does local market type digital work where the images are essentially customized stock photos. The other category it could replace would be brands filling out how an item could look on apparel. For instance, Ralph Lauren doesn't use a model right now for every skew. The other use case for would be interactive such as a customer tweaking the model to match their own skin, eyes, hair, etc. in order to better see how the garment could look at you. Suitsupply already does this with its look builder.