nicce

1. Not storing secrets properly or using hardcoded secrets

2. Wild use of webviews/iframes sometimes easily propagates as XSS in phones

3. Incorrect client-side OAuth 2.0 configuration e.g. with schema-based redirect URLs.

4. Not supporting high-enough API versions, which may prevent some OS-related weaknesses

5. The list is actually very long. Just few top of my mind.

You don’t believe how often people leave secrets in the app or use webviews and iframes badly, misconfigure OAuth in client side and so on. There are many issues where secure API does not help.

This is probably the end now. People don't debate anymore whether they should write only in assembly vs. in C. Sometimes new abstractions emerge and they replace something completely. This time the abstraction was LLM.

> You just install it on your phone and use the app.

Some people on the cybersecurity side are starting to cry....

So it is price increment in the end, if new pro model comes with the new pro price.

On top of that, they don't make all BS statements or malicious tricks used by some unnamed entities.

European users have right to be forgotten. Waiting for the court order to delete all models.

It is a very old idea. But it matters a lot in iPhone’s scale. Many people will have it without buying a separate device.

The only positive side is that it is harder for students to feed university exercises to the agent in cybersecurity and expect it to make them all.

> In keeping with our commitment to user responsibility, following the official launch of V4.1 Flash and prior to the release of V4.1 Pro, all requests to the Pro model will be routed to V4.1 Flash and billed at Flash's price. If you encounter any issues during your comparative testing between V4 Pro and V4.1 Flash, please do not hesitate to reach out to us with your feedback. Thank you for your support!

Wow. Imagine OpenAI/Google/Anthropic doing this! Nope.

That seems to be indeed true. I guess it gets validated quite soon.

That is the point. Someone must verify that the Lean matches the actual theorem, precisely as it should be interpreted.

As long as the proofs itself are correct. How long they were this time?

Edit: at least ~600,000 lines

https://stanfordtechreview.com/articles/openai-buckmaster-na...

Can’t wait to see the human verifying the results and then figure out that the AI model actually cheated and the results are not correct.

I guess the point of this whole forum is "Why not?"

I guess guys from the opposite side would not work there. So that is what will happen more and more.

> They are suddenly very AI obsessed, which is the antithesis of their platform.

Benefits are massive if you have the discipline to stay as master and don't let AI overtake everything and make you lazy.

Imagine a world were you see glasses being sold out just to prove that police is acting against the law.

Unless Codex is vibe coded and it downloads it directly on every installation…

> It's impressive that they're doing such titanic work.

I think it is still a bit too early to say that. The most difficult part is to ensure the security of the implementation, and not just the checkbox compliance against standards. So when this survives the swarm of pentesters, it will be a absolute titanic work. Otherwise it might be too risky to use.

Ouch, already 11 years since the Windows 10 release.

You also want prompt/prefix cache. Otherwise there is a lot of duplicate prefill processing if you fork conversation, have a different chat window or anything like that. It therefore also makes subagents much faster.

>> Fable was the step change for programming > AFAIK: Mistral does not even try to compete in this field.

They have released models speficially for programming that ”vibe coding” would be safer.

https://mistral.ai/news/leanstral/

> Just the other day I saw some meme about how if we saw as many ads taking over sites like we do now, that was a sure sign of having your machine getting infected with a virus.

That used to be the reality around decade ago for those who don’t know. Some applications changed the default front page of the browser, hijacked other websites to inject popups and so on

Better to buy GPUs and RAM with current rate

It is not just the money but social media is controlled by right wing ideologist. Musk has openly supported right wing parties in Europe.

Oops

5h limit is gone with one prompt with Sol or Astra with high or higher thinking.

Well, that makes them just even slower then

That MXFP4 is an excellent project. But I have difficulties on reading that README. Is it intentionally generated like that with LLMs?