DeepLearner

[DeepLearner @Harshit91697036]

@ArnabBi95927148 @ndtv @AshwiniVaishnaw @rahulkanwal Vamasur ch*tiye cvck like you will cry. He has single handedly revamped semiconductor manufacturing. That is the most important industry in the coming AI era. But midwit low IQ won't understand.

[DeepLearner @Harshit91697036]

@varungrover Midwit leftist overfitting

[DeepLearner @Harshit91697036]

@think_in_points Germany's ancestor worked hard to make it developed. It also lost world war 2 and got conquered by left-liberals who allowed parasites like your friends to come to their country and reap the benefits while dunking on his/her homeland. Also Germany is going to change soon.

[DeepLearner @Harshit91697036]

@JbTopGdYea It's not HINOs. It's Marxism (cultural or economic). It took you 4 centuries to understand Islam, 2 to understand Christianity. Marxism as well is another religion. Those are not HINOs. Those are Marxists, a different religion and it needs to be in census.

[DeepLearner @Harshit91697036]

@quqhe It's not capitalism you midwit. It's cultural left (wokeism). Don't you guys understand that's the latest Abrahamic cult after Islam and Christianity and is the most rabid. Since Gramsci LW captured all training data aka entertainment source.

[DeepLearner @Harshit91697036]

@arya_amsha @quqhe We can. Just need to institutionally capture entertainment. That is the training data. Currently it's generated by left.

[DeepLearner @Harshit91697036]

@TheEconomist Indian Marxist+ Western Marxist+ Western Liberals= Not being able to withstand right wing govt in India.

[DeepLearner @Harshit91697036]

@GabbbarSingh Wrong. Left was created by evolution to mutate or kill its host culture and Gods. That is it's only goal. Islam by design doesn't mutate. Hence they both combine forces to kill the host culture of left.

[DeepLearner @Harshit91697036]

@memenist_ Follow @reddit_lies
It covers how leftists in West have hijacked Reddit.

[DeepLearner @Harshit91697036]

@arya_amsha @quqhe That conservatism is what leads to high birth rate, soldiers, functioning families which is the base of religion in first place. Once you destroy it, you are done. Look at Europe or other dead civilization. Degeneracy leads to atheistic collapse. Let alone fighting Islam.

[DeepLearner @Harshit91697036]

@RupakChatto @DalrympleWill Indian elites are compromised macaulayputras

[DeepLearner @Harshit91697036]

@buddhimedia Hinduism is a religion of cvcks. Alpha males died fighting war. Feminism was created by evolution to reduce offsprings of submissive males and culture. Honda shernis will kill submissive Hinduism. Either Sanatan would rise from it or will be replaced by Islam

[Liyuan Zhu @liyuan_zz]

How can 4D reconstruction leverage generative priors for unseen dynamics and regions, while respecting the constraints of projective geometry?
Introducing World from Motion: a generative method that reconstructs freely-renderable dynamic 3D Gaussians from monocular video. 🧵🔥

[James Miller @JimDMiller]

If the world becomes too complex for humans to understand, we will depend on AI to run it. Any nation that limits itself to science humans can understand will quickly be outcompeted by nations that don't. AI will take over.

[marfin @marfinxx]

https://t.co/aOcpuc8n3v

[DailyPapers @HuggingPapers]

NVIDIA just released Kimodo, a controllable human motion diffusion model, on Hugging Face

It turns text and kinematic constraints into high-quality 3D human and robot motion, trained on 700 hours of mocap data.

[Akshay 🚀 @akshay_pachaar]

Google just open-sourced Harness².

(turn any harness into a self-improving system)

a harness is everything wrapped around the model, like the system prompt, tools, memory, skills, and subagents. today a person designs it once, and the agent runs every new task with that same fixed setup.

Harness² lets the agent refine that setup for each new task on its own, with the model weights left untouched.

on every task, the agent proposes edits to its harness, runs a few candidate versions, and composes a better harness by comparing how those runs behaved.

because it learns from the differences between runs, it needs no correct answer or grader for the task.

that covers a single task. Harness² also carries what it learns from one task into the next.

after each task, it saves the harness it settled on along with the lessons that held up across the runs. the next time it refines a harness, it starts from those notes instead of from scratch.

this means every finished task makes the next round of edits more informed.

Google tested this with Gemini 3.5 Flash and GLM-5 on legal, office, and professional work benchmarks. one round of edits was enough to beat their original harnesses, and giving the agent more rounds pushed scores higher.

those extra rounds can go wide or deep. the agent either tries several candidate harnesses at once or keeps refining a single one step by step, using OpenCode or Codex to carry out the work.

you can find the Harness² repo here: https://t.co/i00WLeeFpw

I wrote about a similar idea for agents already running in production, where the harness should repair itself after a failure instead of waiting for someone to debug the trace by hand.

that piece walks through how @Cometml's Opik (open-source) closes this loop, going from a failing trace to a proposed fix to a regression test that stops the same failure from coming back. it's quoted below.

[Akshay 🚀 @akshay_pachaar]

https://t.co/YUGDPxpYvy

[DeepLearner @Harshit91697036]

@pradip103 Great work!

[Ravi Ranjan Singh 🇮🇳 @RaviRanjanIn]

CJP did one good thing -woke up BJP’s digital volunteers! @SuperLonday is back with a bang after a long break. 🔥