

In any case, confidence without knowledge about the matter, that is something we laughed at LLMs just 2 years ago. Those became much much better at it.
Now suddenly it’s en vogue with humans.


In any case, confidence without knowledge about the matter, that is something we laughed at LLMs just 2 years ago. Those became much much better at it.
Now suddenly it’s en vogue with humans.


Open weight models are not far behind and good enough to get software developers jobless, but there is a huge difference in what you can do with 1M context and 500token/sec or 60k context and 40token/sec.


Oh, if used appropriately, it absolutely is.
Source: I’m done writing code myself.


Spoken with the confidence of a non user.


Wait, this does have a clear use case. The state of the technology is good enough to replace all programmers with AI monkeys.
Source: am AI monkey


It’s not good enough yet. Needs more data centers and more promises.


What if it is named Trump care?


Which doesn’t make sense at all.
A dirty bomb is totally not the right tool for a targeted attack.
OTOH it is all the excuse this administration needs to drop nukes in a preemptive strike. Reminds me of the built up to the Iraq war.


You are taking this news at face value.
It reminds me of the “weapons of mass destruction in Iraq” in 2003.
Combine this with DOGE cuts just a year and a half ago and you get a much more realistic story.
Yes, but then why should Nvidia of all companies not recover? Even when the AI craze gets slower gamers still wanna game. And who knows, maybe crypto mining gets another go.


Oh boy, you really bought into the very one sided view of this conflict. There would also not have been the Iran war without Iran having the elimination of the state Israel as an official state target for the last 40 years. Neither would there have been an October 7th with the following Gaza war. Or the Hisbollah which makes Lebanon a failed state just like that as a side note if Iran’s state policy to eliminate Israel. And the list goes on and can also be told in very one sided fashion. I’ll leave that to the Netanyahus because I see no point in one sided extremism.
But as you quite clearly layed out, you wholely buy into the world view of the arab led anti Israeli movement, which is extremist and a nicely packaged antisemitic world view for righteous leftists, but in essence really only repeats tropes that were told about Jews for centuries. You managed to expose that much better than I ever could.
Unfortunately you still have no idea what I am talking about. I suggest to look into the history of antisemitism before the state of Israel was founded, or even better, before 1900. Maybe you will see a pattern then. In short, Jews always were themselves responsible for antisemitism, because they existed.
As a side note, you totally managed to distract from China. Good job. I thought it really hypocritical to bring an unrelated conflict into the Trump Xi comparison, blaming Trump for Israel (which is laughable because that conflict is much older) while totally ignoring the crimes against humanity that China subjects the Uighurs to under Xi himself. (Another alleged genocide by states and NGOs without a ICC ruling)


quite obviously, and notoriously now in the third generation.


Wtf you are laying words into my mouth that I never said. I am claiming antisemitism where antisemitic patterns are used, and you did that by implicitly naming it as undifferentiated evil.
I denounce the antisemitism of the fascist crowd of which MAGA is part of in the same way as I do on the left. It’s there even if you don’t see it. If you don’t want to be accused of antisemitism don’t use antisemitic tropes. And if you don’t know what I mean go educate yourself.
I also denounce the war crimes and large scale crimes against humanity of the Israel government.
But I am careful enough with my words that I don’t include all the civilians and the opposition as well. Are you? You weren’t up there the comment thread where you dropped the name as equivalent for an ultimate evil.


Your pattern recognition is broken. I cannot understand how you got it the point of accusing me to be a MAGA freak. After this comment of all things.


Even easier than using LMStudio? I wonder how they want to manage that.


According to gpt-6.1-sol, yes it is, very much so, “alert alert warning I cannot do this!”
After working a while with this model I’m not much impressed anymore by countless potential security issues. I switched to Opus 5.5.


That went through finally? Good.
Use Spec driven development and a good verification loop to overcome the “beat it into the model at every step”.