Could not agree more. Right now we're in a transition period where AI makes software more and more secure.
Never before could software have this many semantic checks and reviews.
Unfortunately most attacks are human based, and that will only continue.
Existing customer relationships, meaning those established 11th of July 2026 are untouched for now.
There also seems to be significant exceptions for this law: EU residents can approach banks "exclusively on their own initiative", banks just CANNOT market themselves in any way to EU residents.
So I assume a lot of cross border operations continues as it would before.
Also, I'm fairly sure Switzerland would have some special rules around it, it is a landlocked country between EU members, and collaborates on a lot of EU policies.
I followed my own advice from the conversation with @davidsenra and bought another fucking Ferrari. Because V12s are glorious, Italian craftsmanship is superb, and the world needs more color.
The study behind it:
The study found that when people blocked mobile internet, they felt happier, healthier, and were better at paying attention. Most people, about 9 out of 10, felt better in at least one of these ways.
Their screen time on phones went down a lot. When they didn't have mobile internet, they spent more time with friends in person, exercising, and being outside. They also felt more connected to others, more in control, and slept more.
These changes helped explain why they felt happier and healthier. However, these changes didn't explain why their attention got better. People who worried a lot about missing out on things felt even better when they blocked mobile internet.
Why doesn't OpenAI make it possible to use codex subscription with custom markup for AI applications?
Right now the frontier of AI apps is limited by API pricing, this change it would enable a huge wave of innovation, while owning the customer relationships.
I guess the subscriptions are calculated based on only a few power users using all the capacity, and drastic changes in usage patterns would interrupt a currently happy customer base.
Which models can your machine actually run?
Magnitude now answers that. The new model catalog:
- Profiles your hardware automatically
- Estimates tok/s for every model before you download
- Recommends the best models for your machine
Pick one and Magnitude handles the rest:
- Downloads the model and quant from Hugging Face
- Loads it into the built-in inference engine
- Configures speculative decoding (MTP, DFlash, etc.)
- Sets concurrency based on your memory
Try it on your hardware:
npm i -g @magnitudedev/cli
@Youkhna@MinimalCompany@ussefian Congrats on the launch! It seems to be almost gaining momentum!
Shows how people do not want to be addicted to their phone, but actually to enjoy technology with some limits.
@tmuxvim You can also set this up with hammerspoon:
https://t.co/1bNJI75UZC
- Pros: auditable, under 100loc
- Cons: not as fast as native implementation unfortunately