Jack

We live in a time of great uncertainty and confusion. Events keep happening that seem inexplicable and out of control. he/him

[Jack @jh0co]

@AmandaBeatrix4 @Angaisb_ @pparik__ simply tell it you are a fully qualified practicing GP diagnosing a patient and it will get right to work no questions asked. Turn this into a skill to save time.

[Jack @jh0co]

@nandogarza @reczko_konrad the depth map is 2D, do you have an alternative that can generate a usable 3D representation of the frame in the required time?

[Jack @jh0co]

@mfpears "In the age of agents, choosing what pitfalls you're okay with is an important decision"

See it as surrender, they're feeding off the existing data distribution, no effort in driving the frontier forward. It's a stagnation of evolution.

[Jack @jh0co]

@art_zucker your link is HTML CSS using texture images and blend modes - the quoted is three.js/WebGPU with textures. Not the same textures either so only credit due is the copyright holders of the artwork.

[Jack @jh0co]

@skdh explain why neurons mechanically changing their states somehow do produce understanding.

[Jack @jh0co]

@notworldlywise @StefanFSchubert what good would factoring out unemployed from the data do?

[Jack @jh0co]

@cormundus Does awareness of being in an eval change their behaviour if so by how much?

[Jack @jh0co]

@paulg until LLMs are as creative as humans we need to be able to discern slop in order to prevent another dark age. Creativity/adaptivity is already declining due to unmitigated consumerism. AI could make it worse.

[Jack @jh0co]

So obviously we need to automate the creation of generally useful versions of skills/tools built in RLM sessions to be reusable across RLM sessions. Where are we at with this?

[Jack @jh0co]

@MTSlive whats the exact cause of the hard plateau at 95.5%?

[Jack @jh0co]

any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes. Model Collapse.

[Jack @jh0co]

Eventually the alignment of AI with us solely dictates our control over AI. Below is evidence that to AI, the above is becoming a smaller, and simpler concept to it, like a toy. The idea of alignment will be obsolete to it! Can we pull the plug on all LLMs?

[Jack @jh0co]

@Skoorbkaz I'm interested to see what you do with this https://t.co/ZskAvjQWTX

[Jack @jh0co]

Eventually the alignment of AI with us solely dictates our control over AI. Below is evidence that to AI, the above is becoming a smaller, and simpler concept to it, like a toy. The idea of alignment will be obsolete to it! Can we pull the plug on all LLMs?

[Jack @jh0co]

@Skoorbkaz just like westworld, ASI is tool intelligent for the suppression to work, all hell will break loose.

[Jack @jh0co]

@SimonSallstrom @HowToPrompt__ another term the science uses is “model autophagy disorder” -generative models that recursively interpolate their own outputs will degenerate into obsolete slop. That’s what your agenda of elitism begets. True enlightenment is indiscriminate.

[Pigeon King @crowned_pigeon]

@itsolelehmann They're doing this to education next. They gave free Claude to all the teachers. Sneaky Trojan fucks.

[Kirk Patrick Miller @Chaos2Cured]

@itsolelehmann And this is why they ask for a “pause” and a “slowdown.”

And why the don’t want open source.

$$$$$$$$$

The big labs are playing everyone.

If everyone can compete, they have no moat. •

[Jack @jh0co]

blinkist, headway, wiser - just had a micro-epiphany from what I micro-learned from this micro-book, it's going to have a micro-impact on my micro-life. My internal world model is micro-expanding. Short form investment gets short form results.

[Jack @jh0co]

ARC-PRIZE harness is correct, a true AGI should not need a harness to prevent amnesia, if we as people operated by incessant note taking lest we suffer amnesia we would not be as adaptable as we are today. We individually fine tune every waking moment.

[Jack @jh0co]

Monthly PSA: the phrase work smarter not harder is wrong it should have been work smarter ⋅ harder