karol

[“paula” @paularambles]

You see, what fascinates me about “remember to not smile” is that he needs his pain to be witnessed. He suffers at 4am, okay, fine, many people suffer. But then, immediately, he makes a slide! He gathers the entire company! He writes an article! Is it not clear what is happening here? Suffering without an audience simply does not count. This is ideology at its purest.

We see here the old Protestant trick in its Silicon Valley form: success alone is vulgar, almost pornographic. My god, you cannot simply succeed. It must first be purified through unnecessary suffering. So you get this obscene reversal where pain stops being a cost of the product, and instead becomes the product itself. The developer documentation is, how do you say, merely the byproduct, the excrement of the true production, which is the suffering. As Hegel already knew, and here I think even Lacan would agree, although perhaps not, the true obsession is always with deserving success.

Here I must tell you this old joke from Soviet Union. A worker says: "we pretend to work, they pretend to pay us.” Now here it is inverted, which I claim is much worse. They are paid perfectly well, wonderfully well, and so they must pretend to suffer. Do you see the perfect madness of this?

And now, the ultimate obscenity. Imagine the company succeeds while everyone leaves at 5pm. Catastrophe! Total ontological catastrophe! In this worldview, I am tempted to say, failure would almost come as a relief. Failure means you simply did not suffer enough, go back, suffer more. But success at 5pm? This cannot be explained.

So when he says "remember to not smile," this is a desperate command of the superego. If you smile, we will all discover the pain was never necessary.

[du @thedulab]

If you're aware that the term "permanent underclass" even exists then it's clear you'll never be a part of it. No shot. On this app all day exposing yourself to every opinion about what the future might look like

You've gotten to a point where you've essentially become an expert in predicting exactly how you might be cooked. With that level of critical thinking there's just no reality where you'll ever allow it to happen to you. Even seeing this on your feed is already proof of that

Seriously though you have to give yourself some more credit. You're not some useless sack of flesh and you never were. The fact that this stuff is even on your mind means you care enough about your life to never give up

Keep moving forward and trust in your ability to adapt to whatever life throws at you. No need to overly fear because you'll always be able to figure it out. Impossible not to figure out out being surrounded by so many bright minds. Believe it or not you're one of them

[Ronak Malde @rronak_]

A cool lesser known fact in LLM history -

Every lab does their tool calling format very differently, and for the longest time this was a pretty important secret to be kept.
Nous Research was the first to open source their Hermes tool calling format and model, which crushed the BFCL tool calling leaderboard back in 2024, and showed people a reliable tool format and parser.

Right after, Qwen started using the Hermes tool format too, and now all Qwen models still use it to this day. The Hermes format also became the bedrock for our swe-1 models at Windsurf.

Really cool to see the contributions of this lab live on in open source.

Here's what it looks like:
<|im_start|>assistant
<tool_call>
{name: args:}
</tool_call>
<|im_end|>

[Dave 🧊 @DavidCarlson]

Our country was founded by sensitive young men. Not wife guys. Remember.

[karol @zioutens]

@ryder_ripps Aren’t you that guy that was writing articles about how Yuga Labs are neonazis and BAYC is full of dog whistles? Why do you own one all of a sudden

[karol @zioutens]

@civil_capybara @luciascarlet This already exists, notifications can be sent as urgent

[karol @zioutens]

@viemccoy

[karol @zioutens]

@f4micom How do you know it's AI? Maybe just a retarded designer

[karol @zioutens]

@daisyui_ The snippet he provided won't even work because the second &lt;ul&gt; won't have the styles applied

[karol @zioutens]

@mattpocockuk Plan mode using Opus and exection using Sonnet, or both Opus?

[karol @zioutens]

@0xKyon I mean it's easier to be stoic when you're drugged with opium all the time as Marcus Aurelius was

[karol @zioutens]

@_opencv_ People in the comments missing the point. This post is not saying anything about the merit of the statement, just pointing out LLM brainrot speak "not x, but y"

[karol @zioutens]

@airkatakana I guess I'm lucky because I paid for a year in advance and don't have these weekly limits at all

[𓏴𓀿𓁼𓀐𓏴 @404vish]

there's a non-disparagement clause i was asked to sign by my forbes 30u30 co-founder when she left the company while i was still healing from the blood clots in my brain 🤷‍♂️ so i can't say anything at all

[karol @zioutens]

@Yuchenj_UW For the phone number extraction - there might cases where there are multiple numbers in the email and then you might need to parse relevant context to decide which one you want to get. People also might format them in different ways and it's hard to cover 100% of cases

[karol @zioutens]

@Yuchenj_UW For some cases I agree, however for some tasks like profanity filter an LLM is actually the better way to do it because it's much harder to bypass (for example by using special characters, a different language etc.)

[★ @thematrixwizard]

Gm 🌞

[karol @zioutens]

@Shpigford I use https://t.co/55fp4RHg86

[derek guy @dieworkwear]

I respectfully disagree. To me, this represents a poor financial decision. Let me show you why. 🧵

[karol @zioutens]

@BonesawMD Maria Skłodowska-Curie