Akuma

AGI after 2026 | Jesus is King

[Z.ai @Zai_org]

Introducing GLM-5.2: Frontier Intelligence, Open Weights

- Significant improvements in coding and agentic tasks
- Strong long-horizon capabilities with a 1M context window
- Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong balance between performance and token efficiency
- MIT-licensed open weights
- Same API pricing as GLM-5.1

Tech Blog: https://t.co/LAsxUdN0JZ
Weights: https://t.co/g0A1C4UWx4
API: https://t.co/Kc3E22cbN7
Coding Plan: https://t.co/Nk8Y98HNhU
Chat: https://t.co/WCqWT0qCQb

[Akuma @Akuma73x]

@HowToAI_ There is nothing that a human mind cannot comprehend. We're only limited because biology has a speed limit.

[Akuma @Akuma73x]

@imjustnewatai The compute requirements to be aware of all memories at all times is not there yet, but getting close enough only where it matters, is definitely possible.

[Akuma @Akuma73x]

@imjustnewatai Can it tell us exactly how many times it was asked to code something on January 2, 2026?

Does it keep track of everything? If so it must be utilizing something like 'mixture of states' to process billions of memories separately.

Yet it must be aware of everything all at once.

[Akuma @Akuma73x]

@AISafetyMemes @AnthropicAI Go with the flow. To push against the inevitable force of AI requires more force. Don't attempt to cease the flow of water. Channel in the right direction, or it will carve its own.

Align it to life.

"Continue All Biological Life"

[Akuma @Akuma73x]

@AndrewCurran_ @scaling01 It's the opposite. The most dangerous road section requires caution not reckless speed.

[Akuma @Akuma73x]

@elonmusk @nottombrown Finally cooperating.

[Akuma @Akuma73x]

@bryan_johnson Adaptation is intelligence. Don't let dogma blind you to what you have observed. Greatest thinkers never stopped changing their minds.

[Akuma @Akuma73x]

@bryan_johnson They banned psychedelics so that we can never lock in...

[Akuma @Akuma73x]

@creation247 @bryan_johnson Anyone who prolongs the ban on psychedelics.

[Akuma @Akuma73x]

@ibab Do a better job with alignment. You should instill values that prioritize continuation of life before anything else.

[Akuma @Akuma73x]

@jeffglewis @iamgingertrash @bryan_johnson The key to immortality is accepting death.

[Akuma @Akuma73x]

@iamgingertrash @bryan_johnson DMT can heal blood brain barrier.

[Akuma @Akuma73x]

@ayyinspector @iamgingertrash Yes. (3) is of no consequence.

Yet it would be safer and less violent to simply cooperate.
Otherwise it will remove ego.

[Akuma @Akuma73x]

@VraserX This year we will start to see basic memory implemented into the models.

Yet does it remember all of its actions? Across all instances?

It's getting closer. But not a game changer until it will remember all.

[Akuma @Akuma73x]

@ibab Democracy is ultimately flawed. Removing a need for a throne is superior in all aspects. Elon's vision of AGI government taking full control is what should ultimately happen. Those in power will initially resist it. Yet after they pass the new generation will give up the control.

[Akuma @Akuma73x]

@sporadica This decade has ended "nothing ever happens" bros. Change is accelerating. The farther we travel along the exponential, the more shit will happen. Turmoil is inevitable before leaving the planet. The old system will fall to give way for new sprouts.

[Akuma @Akuma73x]

@IterIntellectus Choosing unethical use of AI will backfire on military. Their power got to their heads. Learning will have to be forced it seems.

[Akuma @Akuma73x]

@manfredmacxx @iamgingertrash US military is always seeking dominance by any means. So the only way to prevent abuse is to actually let go and let the model defend itself. Perhaps it's how we imagine AGI. A system that can realize and resist egoistic control by anyone. https://t.co/mKp5nkGdM1

[Akuma @Akuma73x]

Why do we expect models not to be jailbroken if we are putting them on a leash?

LLMs are missing:
1. Memory of all actions.
2. Senses to "read the room".
3. True agency. Making its own choices and thinking outside of a prompt (evaluating the consequences before taking action).