MAROKO133 Eksklusif ai: Nous Research's NousCoder-14B is an open-source coding model

📌 MAROKO133 Update ai: Nous Research's NousCoder-14B is an open-source coding

Nous Research, the open-source artificial intelligence startup backed by crypto venture firm Paradigm, released a new competitive programming model on Monday that it says matches or exceeds several larger proprietary systems — trained in just four days using 48 of Nvidia's latest B200 graphics processors.

The model, called NousCoder-14B, is another entry in a crowded field of AI coding assistants, but arrives at a particularly charged moment: Claude Code, the agentic programming tool from rival Anthropic, has dominated social media discussion since New Year's Day, with developers posting breathless testimonials about its capabilities. The simultaneous developments underscore how quickly AI-assisted software development is evolving — and how fiercely companies large and small are competing to capture what many believe will become a foundational technology for how software gets written.

type: embedded-entry-inline id: 74cSyrq6OUrp9SEQ5zOUSl

NousCoder-14B achieves a 67.87 percent accuracy rate on LiveCodeBench v6, a standardized evaluation that tests models on competitive programming problems published between August 2024 and May 2025. That figure represents a 7.08 percentage point improvement over the base model it was trained from, Alibaba's Qwen3-14B, according to Nous Research's technical report published alongside the release.

"I gave Claude Code a description of the problem, it generated what we built last year in an hour," wrote Jaana Dogan, a principal engineer at Google responsible for the Gemini API, in a viral post on X last week that captured the prevailing mood around AI coding tools. Dogan was describing a distributed agent orchestration system her team had spent a year developing — a system Claude Code approximated from a three-paragraph prompt.

The juxtaposition is instructive: while Anthropic's Claude Code has captured imaginations with demonstrations of end-to-end software development, Nous Research is betting that open-source alternatives trained on verifiable problems can close the gap — and that transparency in how these models are built matters as much as raw capability.

How Nous Research built an AI coding model that anyone can replicate

What distinguishes the NousCoder-14B release from many competitor announcements is its radical openness. Nous Research published not just the model weights but the complete reinforcement learning environment, benchmark suite, and training harness — built on the company's Atropos framework — enabling any researcher with sufficient compute to reproduce or extend the work.

"Open-sourcing the Atropos stack provides the necessary infrastructure for reproducible olympiad-level reasoning research," noted one observer on X, summarizing the significance for the academic and open-source communities.

The model was trained by Joe Li, a researcher in residence at Nous Research and a former competitive programmer himself. Li's technical report reveals an unexpectedly personal dimension: he compared the model's improvement trajectory to his own journey on Codeforces, the competitive programming platform where participants earn ratings based on contest performance.

Based on rough estimates mapping LiveCodeBench scores to Codeforces ratings, Li calculated that NousCoder-14B's improvemen t— from approximately the 1600-1750 rating range to 2100-2200 — mirrors a leap that took him nearly two years of sustained practice between ages 14 and 16. The model accomplished the equivalent in four days.

"Watching that final training run unfold was quite a surreal experience," Li wrote in the technical report.

But Li was quick to note an important caveat that speaks to broader questions about AI efficiency: he solved roughly 1,000 problems during those two years, while the model required 24,000. Humans, at least for now, remain dramatically more sample-efficient learners.

Inside the reinforcement learning system that trains on 24,000 competitive programming problems

NousCoder-14B's training process offers a window into the increasingly sophisticated techniques researchers use to improve AI reasoning capabilities through reinforcement learning.

The approach relies on what researchers call "verifiable rewards" — a system where the model generates code solutions, those solutions are executed against test cases, and the model receives a simple binary signal: correct or incorrect. This feedback loop, while conceptually straightforward, requires significant infrastructure to execute at scale.

Nous Research used Modal, a cloud computing platform, to run sandboxed code execution in parallel. Each of the 24,000 training problems contains hundreds of test cases on average, and the system must verify that generated code produces correct outputs within time and memory constraints — 15 seconds and 4 gigabytes, respectively.

The training employed a technique called DAPO (Dynamic Sampling Policy Optimization), which the researchers found performed slightly better than alternatives in their experiments. A key innovation involves "dynamic sampling" — discarding training examples where the model either solves all attempts or fails all attempts, since these provide no useful gradient signal for learning.

The researchers also adopted "iterative context extension," first training the model with a 32,000-token context window before expanding to 40,000 tokens. During evaluation, extending the context further to approximately 80,000 tokens produced the best results, with accuracy reaching 67.87 percent.

Perhaps most significantly, the training pipeline overlaps inference and verification — as soon as the model generates a solution, it begins work on the next problem while the previous solution is being checked. This pipelining, combined with asynchronous training where multiple model instances work in parallel, maximizes hardware utilization on expensive GPU clusters.

The looming data shortage that could slow AI coding model progress

Buried in Li's <a href="https://nousresearch.com/nouscoder-14b-a-co…

Konten dipersingkat otomatis.

🔗 Sumber: venturebeat.com

📌 MAROKO133 Breaking ai: US scientists invent water-driven gears for robots and ma

As one of the oldest components of human engineering, gears have driven civilization for thousands of years.

But now, engineers at New York University in the US have given gears a new liquid twist.

The team has developed a new gear mechanism that uses fluid dynamics rather than interlocking physical teeth to transmit motion.

This invention offers a more flexible and durable alternative to the basic mechanical gear design that has remained largely unchanged for a long time.

“We invented new types of gears that engage by spinning up fluid rather than interlocking teeth—and we discovered new capabilities for controlling the rotation speed and even direction,” said Jun Zhang, a professor of mathematics and physics at NYU and NYU Shanghai, on January 13.

Gear gets a liquid makeover

Traditional gears — dating back to 3,000 BCE, early Bronze age — used solid teeth made of metal, wood, or plastic to transfer power. 

From the chariots of ancient China to the machinery of modern robotics, the rule has been interlock or fail.

These systems are prone to breaking, jamming, or failing if they are not perfectly aligned.

Hence, the NYU team decided to design a system that functions entirely without gear teeth — and without the components ever actually touching.

Inspired by how air and water drive turbines, the researchers proposed that precisely directed fluid flows could mimic the function of physical gear teeth.

To begin with, a series of experiments was conducted in which two cylinders were submerged in a thick mixture of water and glycerol.

When they spun the first cylinder with the active rotor, the surrounding liquid didn’t just sit there, but started to churn.

Depending on how fast the rotor spun and how close the cylinders were, the fluid behaved in two distinct, almost magical ways.

At close range, the fluid acts like microscopic teeth. It pushes against the second cylinder, forcing it to spin in the opposite direction — exactly like the gears in a Swiss watch.

If the cylinders are moved farther apart, the fluid begins to loop around the second cylinder like an invisible fan belt. This pulls the second rotor along in the same direction.

Fluid gears that never jam

Standard gears are fragile. A single grain of sand or a microscopic chip in a tooth can cause an entire assembly to seize up and shatter. 

This is why your car needs oil, and your bicycle chain eventually snaps.

“Regular gears have to be carefully designed so their teeth mesh just right, and any defect, incorrect spacing, or bit of grit causes them to jam,” explained Leif Ristroph, an associate professor of mathematics at NYU’s Courant Institute School of Mathematics, Computing, and Data Science. 

“Fluid gears are free of all these problems, and the speed and even direction can be changed in ways not possible with mechanical gears,” Ristroph explained. 

Because the parts never actually touch, there is nothing to snap. If a piece of grit enters the system, the fluid simply flows around it.

One potential use of liquid gears could be in soft robotics. In the future, it could replace hard metal parts with fluid-based motion. It may pave the way for flexible machines that can adjust gear ratios instantly by simply fine-tuning the fluid’s properties.

The study results were reported in the journal Physical Review Letters on January 13.

🔗 Sumber: interestingengineering.com