OpenAI cracked these frontier math proofs using models and amount of compute not available to general public. It's not the case that anyone can put "solve Navier–Stokes" in ChatGPT and get solution. Although less well-known problems can be solved with publicly available models.
Indeed. Their model they showed off that was ~5x better than 6 Astra in certain domains including mathematics is not public and I doubt they'd release it. They distill their internal large into their public models.
Not only are the models and compute unavailable to the public, they are unavailable to anyone besides the labs, including the government. We better hope the furthermost tip of the jagged frontier stays in purely theoretical mathematics, because if it strays into the physical world then we may find ourselves in a situation where the power of the labs rivals the governments supposedly overseeing them.
Latest LLMs that have good visual understanding (Gemini and GPT) are able to tell if construction work is up to code. You can give it a picture of electrical wiring, plumbing, framing and it will tell you if it's up to code with very good accuracy.
True. Depending on how heavily they compressed it to fit between those studs it will likely be less than R13. No matter whether you use fiberglass or rockwool, you should make sure you buy insulation sized to the depth of the cavity it will insulate. Then use an insulation knife to cut in for obstacles within the cavities.
The current version of FSD drives better than most humans. You literally put in the destination address, press Start Self Drive, and it gets to your destination without you ever touching the steering wheel, and parks for you in the end. IMO it's ready for large scale robotaxi deployment. I only take steering wheel if I want to take a different route or park in a different place vs where Tesla decides to park.
Uber rides are around $1.5 per mile on average. Assume Waymo car drives 200K miles over it's lifetime, 3mi/kWh EV efficiently and $0.20/kWh electricity cost. It will generate $300K over it's lifetime and use $13K worth of electricity. Right now it costs $150K to manufacture Waymo car because they are custom built and Lidars are still expensive. Even at $150K unit economics should be OK. But the cost of self driving cars will go down as they are manufactured at larger scale and will eventually approach cost of regular mid-range car. In a few years they will be under $50K per car. Mapping is a one-time cost and the need for mapping will eventually disappear. This will be a massively profitable business for self-driving fleet operators.
Good calculation approach. Waymos will drive a lot more than 200K miles in their lifetime. Taxis often do a million miles. There are already photos out there of Waymo odometers over 100K.
Take-or-pay contracts appear as "contractual commitments" in 10-K. They are not hidden. That's the way they are reported in all industries where take-or-pay contracts exist. There's nothing nefarious about it.
Yes, the contractual obligation shows up, but not the debt the subsidiaries took on. You have to follow the entity structure and look at their books for that piece. And it’ll show up as a different accounting category, “services rendered,” for instance, not as debt repayments.
