If the price is expected to fall over time, then the negotiated price is below market at the beginning and above market at the end. From the suppliers view, they take a loss in the early years that they recover later.
Apple probably pays a premium to shift this risk to the supplier. Besides that they don't take a loss just because prices tend to fall. They only lose if market prices fall more steeply than expected.
OpenGL doesn't have any way to do this except sometimes via vendor specific extensions. Basically how OpenGL works is it creates the graphics context on whichever device the system hands it. So you can configure the GPU used by OpenGL on the system level but not at the application level.
FYI, setting `NvOptimusEnablement` and `AmdPowerXpressRequestHighPerformance` have been the canonical way on Windows, `__NV_PRIME_RENDER_OFFLOAD` + `__GLX_VENDOR_LIBRARY_NAME` on Linux. Though not an OpenGL feature per se, as you mentioned.
I had a coworker who spent his career working for 2 years then taking 2 years off to live off the savings. Have you considered such a strategy? It seemed to be working out well enough for him.
Or maybe its all about compute availability like they say. It could be that they plan to start training a new model on the 22ed, so the amount of compute available for inference will be greatly reduced.
## A new data retention policy
Finally, we’re making a change to the way we handle business
customer data for Fable 5, Mythos 5, and future models with
similar or higher capability levels. We will require 30-day
retention for all traffic on Mythos-class models, on both
first- and third-party surfaces. [...]
No it says sharing is required. If you don't change the setting on your account then you simply can't access Fable, its not like the setting is ignored. I just tried this on my account and it blocks API requests to Fable.
Interesting. When I tried switching to it yesterday from Opus 4.8 ("/model" command) it didn't complain, but I didn't actually send anything to it when I saw the cost was like 2x Opus 4.8. ie switched back
I'll try to remember to actually try it tomorrow and see what happens.
They are encrypted to prevent others from training on the reasoning trace. But before anthropic started encrypting the reasoning traces they were signed. The signature was to prevent the user from being able to manipulate the reasoning part of the context because this reasoning part is considered "more trustworthy" by the model. So it would be bad if the user could manipulate the reasoning to convince the model to do something dangerous/against policy. Encryption keeps this property while also preventing the reasoning from being used for training.
The state of incorporation doesn't often matter in interstate contracts. Usually they will include a choice of law and forum selection clause specifically to avoid this kind of thing.
look into "sealed sender" schemes, they allow the recipient to verify the sender identity without allowing anyone else to verify the sender identity. So making use of such a scheme allows you to prevent governments or corporations from intercepting even the metadata of communications (and of course E2EE prevents intercepting the contents).
reply