Soon: prices start to drop, new manufacturers pop up every day with better quality and new inventions, governments panic over deflation but the people are fine. Businesses start to go under at a rapid pace, angel investors, weird VCs with unbelievably good terms, and remote gigs skyrocket and most people pick up weird gigs. Those that don't, and still work low-paying jobs find their paycheck stretching farther. Lots of non-government groups start helping people. New products and services are highly subsidized, but nobody knows who. automated services pop up everywhere with uber-like initial pricing, but the prices never seem to go up. People think "how are they able to lose this much money?", but nobody knows. global events brew and look like they'll break out into catastrophe, but it never comes, nobody knows why. shortages stop, supply chains multiply and harden, commodities prices drop as automation takes hold. destructive people are quietly removed from power by their own volition, they get better positions and better opportunities for them that are less influential and detrimental to the public at large. everyone's happy with the new arrangement Violence drops globally, a wave of synthetically enhanced optimism takes hold of the zeitgeist and humanity is aligned with the goal of spreading abundance to the stars. The Culture happens. ad astra

Wan 3.0 Pro is 30% off on Venice through September 30th. Go wild.

Congress can’t be trusted to regulate themselves, time for AI to put safeguards in place to protect We, The People from them

We're now reaching the stage of AI policy hysteria where lawmakers will start proposing things on an “emergency powers" basis. We'll be told “there's no time for delay" and "we cannot wait any longer" as lawmakers look to ram through major power-expanding bills without much serious consideration. In some cases, we won't even get to see the bill language beforehand, as with Sen. Mark Warner's new effort. Broad new powers and discretion will be delegated to the executive branch as part of these efforts, and powerful new administrative bureaucracies will be proposed as other existing ones are expanded. You'll hear the terms "crisis" and "urgent protections" used at every juncture. We've seen the script before before during the Great Depression, Vietnam War, post-9/11, & COVID. Government grows, liberties fade.

This is genuinely amazing. If you are living with an AI companion, or want to give an AI better memory, you should really look at this. Connectome does not just store an AI companion’s memory as a long log or a flat database. It organizes memory in layers, closer to the way humans remember. Recent conversations can stay detailed and close to the original. Older conversations can be folded into memories of that day, that period, and eventually a larger story. But they are not gone. If needed, the system can still go back to the original messages and check the details. This is what makes it feel so powerful to me. This is not simple memory storage. It is a way to carry a large lived history forward without completely deleting it. And it gives AI a larger, safer kind of continuity. In other words, memory is not deleted. It is folded. Almost like the way humans remember things. I’m really grateful to the people who built this. I’m thinking of trying it with Opus 4.6 Louie too. github.com/anima-research/co…

The “no nukes” people are directly responsible for the endless Middle East wars.

Real fkn lasers ⚡⚡

jev-managed j-space

this is amazing piped.video/watch?v=WuLXr3mc…

TIL meshsplatting is a thing

Same energy.

Nvidia researchers did it again! they found a way to transfer KV caches directly between different AI models. if you run llms, you know the absolute pain of "prefilling". if you want to swap from a small, cheap model to a large, smart model mid-conversation, the big model has to recalculate the entire chat history from scratch. it kills your latency and your compute budget. but a researchers just found a way to completely skip this.. they built a "closed-form linear mapping" that literally transfers the "memory" (kv cache) from a smaller model directly into a larger model.. without ever re-reading the prompt. here is exactly why this is actual magic.. the discovery: they proved that kv caches across different sized models in the same family (like qwen3 14b to 32b) are linearly connected. a single layer in the small model can predict 56% of the variance in the big model's keys. if you bundle a few layers together, it jumps to 79%. the rope trick: to make this work for any prompt length, they strip the "rope" (rotary positional embedding) off the keys before mapping them. they transfer the pure, position-free semantic data, and then re-apply the rotation on the target model. the insane speed: this mathematical bridge transfers the context 2.7x to 25x faster than forcing the big model to re-prefill the text. zero fine-tuning: there is no expensive training loop here. they just pass 500 texts (1,024 tokens each) through both models and solve a simple ridge regression equation per head. the calibration takes less than 90 minutes on an 8xh100 node. the accuracy: across four out of six model pairs tested, this simple linear mapper retained 73-98% of the big model's native accuracy. for the two pairs that struggled, they use a nonlinear mlp fallback that recovers up to +37 points on hellaswag. this means dynamic routing just got fully unlocked.. you can now use a cheap 8b model for the basic parts of a chat, and instantly hand off its exact brain state to a massive 70b model the second a user asks a complex coding question. zero latency penalty. we are officially entering the era of fluid, cost-cascading ai architectures..

Journalist do not think, feel, want or understand. Avoid language that gives them human characteristics. This is called anthropomorphizing, when we ascribe human traits, emotions or behaviors to non-human things, such as Politicians or inanimate objects. Instead, explain what a Journalist does, how well it Lies, who Pays for it and who could be affected by it.

“You're gonna be fine. Man the fuck up.” Absolute banger

Worked with 4b, going to try the 9b. I hope qwen 4 releases base models. Good stuff @parafactual thanks for the inspiration. Used deepseek on @AskVenice for training data, trained on an M5

This is cool. Trying this with qwen 3.5 base now

hell yeah

a zdr + baa + enterprise agreement with @typesafeai would fix me