MAROKO133 Breaking ai: Amazon Admits Its Flagship AI Coding Tool Isn’t Good Enough for Its

📌 MAROKO133 Update ai: Amazon Admits Its Flagship AI Coding Tool Isn’t Good Enough

In November, Amazon leaders sent an internal memo to employees, pushing them to use its in-house code generating tool, Kiro, over third-party alternatives from competitors.

“While we continue to support existing tools in use today, we do not plan to support additional third party, AI development tools,” the memo read, as quoted by Reuters at the time. “As part of our builder community, you all play a critical role shaping these products and we use your feedback to aggressively improve them.”

It was an unusual development, considering the tens of billions of dollars the e-commerce giant has invested in its competitors in the space, including Anthropic and OpenAI. Both of these companies have been caught in a heated head-to-head race to establish dominance in the quickly growing AI coding field — while seemingly leaving Amazon’s Kiro long behind.

Half a year later, Amazon is singing a dramatically different tune. As Business Insider reports, Amazon is officially throwing in the towel, succumbing to growing calls among employees for access to OpenAI’s Codex and Anthropic’s Claude.

The decision highlights how desperate AI companies’ desire to maintain competitive edge — and give themselves the best chance of saving themselves from financial ruin — has become. It’s particularly awkward for Amazon, which has deep ties with several other key players as part of a cloud-driven, hyper-scaling strategy.

That’s not to mention its own doubling down on AI coding tools backfiring spectacularly, with Amazon admitting recent outages were related to poorly implemented AI-generated code.

In a note to staffers obtained by BI, VP of Amazon software builder experience Jim Haughwout announced Claude Code would be made available, with Codex following next week.

It’s not a complete capitulation. Both coding tools will run on Amazon’s Bedrock, a fully managed Amazon Web Services-based software that provides secure access to frontier AI models. But it does feel like a certain admission that the company’s own flagship coding tool isn’t up to snuff compared to the competition.

“To help you invent more for customers, we are expanding the agentic Al tools available to you,” Haughwout told employees.

Earlier this year, software developers at the company had grown frustrated over limitations Amazon had put on the use of Claude Code, as detailed in the November internal memo. Some said it was awkward to promote the use of Claude Code through AWS Bedrock while not being able to use it themselves at work.

“Customers will ask why they should trust or use a tool that we did not approve for internal use,” one employee wrote in a comment obtained by BI.

Meanwhile, given the unfortunate optics of opening the floodgates for Codex and Claude Code, an Amazon spokesperson told the publication in a statement that teams are still “primarily using” Kiro, claiming that 83 percent of engineers at the company are leaning on it.

More on Amazon: Amazon Admits Extensive AI Use Is Wreaking Havoc on Its Core Business

The post Amazon Admits Its Flagship AI Coding Tool Isn’t Good Enough for Its Own Workers to Use appeared first on Futurism.

🔗 Sumber: futurism.com

📌 MAROKO133 Breaking ai: ByteDance Introduces Astra: A Dual-Model Architecture for

The increasing integration of robots across various sectors, from industrial manufacturing to daily life, highlights a growing need for advanced navigation systems. However, contemporary robot navigation systems face significant challenges in diverse and complex indoor environments, exposing the limitations of traditional approaches. Addressing the fundamental questions of “Where am I?”, “Where am I going?”, and “How do I get there?”, ByteDance has developed Astra, an innovative dual-model architecture designed to overcome these traditional navigation bottlenecks and enable general-purpose mobile robots.

Traditional navigation systems typically consist of multiple, smaller, and often rule-based modules to handle the core challenges of target localization, self-localization, and path planning. Target localization involves understanding natural language or image cues to pinpoint a destination on a map. Self-localization requires a robot to determine its precise position within a map, especially challenging in repetitive environments like warehouses where traditional methods often rely on artificial landmarks (e.g., QR codes). Path planning further divides into global planning for rough route generation and local planning for real-time obstacle avoidance and reaching intermediate waypoints.

While foundation models have shown promise in integrating smaller models to tackle broader tasks, the optimal number of models and their effective integration for comprehensive navigation remained an open question.

ByteDance’s Astra, detailed in their paper “Astra: Toward General-Purpose Mobile Robots via Hierarchical Multimodal Learning” (website: https://astra-mobility.github.io/), addresses these limitations. Following the System 1/System 2 paradigm, Astra features two primary sub-models: Astra-Global and Astra-Local. Astra-Global handles low-frequency tasks like target and self-localization, while Astra-Local manages high-frequency tasks such as local path planning and odometry estimation. This architecture promises to revolutionize how robots navigate complex indoor spaces.

Astra-Global: The Intelligent Brain for Global Localization

Astra-Global serves as the intelligent core of the Astra architecture, responsible for critical low-frequency tasks: self-localization and target localization. It functions as a Multimodal Large Language Model (MLLM), adept at processing both visual and linguistic inputs to achieve precise global positioning within a map. Its strength lies in utilizing a hybrid topological-semantic graph as contextual input, allowing the model to accurately locate positions based on query images or text prompts.

The construction of this robust localization system begins with offline mapping. The research team developed an offline method to build a hybrid topological-semantic graph G=(V,E,L):

• V (Nodes): Keyframes, obtained by temporal downsampling of input video and SfM-estimated 6-Degrees-of-Freedom (DoF) camera poses, act as nodes encoding camera poses and landmark references.
• E (Edges): Undirected edges establish connectivity based on relative node poses, crucial for global path planning.
• L (Landmarks): Semantic landmark information is extracted by Astra-Global from visual data at each node, enriching the map’s semantic understanding. These landmarks store semantic attributes and are connected to multiple nodes via co-visibility relationships.

In practical localization, Astra-Global’s self-localization and target localization capabilities leverage a coarse-to-fine two-stage process for visual-language localization. The coarse stage analyzes input images and localization prompts, detects landmarks, establishes correspondence with a pre-built landmark map, and filters candidates based on visual consistency. The fine stage then uses the query image and coarse output to sample reference map nodes from the offline map, comparing their visual and positional information to directly output the predicted pose.

For language-based target localization, the model interprets natural language instructions, identifies relevant landmarks using their functional descriptions within the map, and then leverages landmark-to-node association mechanisms to locate relevant nodes, retrieving target images and 6-DoF poses.

To empower Astra-Global with robust localization abilities, the team employed a meticulous training methodology. Using Qwen2.5-VL as the backbone, they combined Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO). SFT involved diverse datasets for various tasks, including coarse and fine localization, co-visibility detection, and motion trend estimation. In the GRPO phase, a rule-based reward function (including format, landmark extraction, map matching, and extra landmark rewards) was used to train for visual-language localization. Experiments showed GRPO significantly improved Astra-Global’s zero-shot generalization, achieving 99.9% localization accuracy in unseen home environments, surpassing SFT-only methods.

Astra-Local: The Intelligent Assistant for Local Planning

Astra-Local acts as the intelligent assistant for Astra’s high-frequency tasks, a multi-task network capable of efficiently generating local paths and accurately estimating odometry from sensor data. Its architecture comprises three core components: a 4D spatio-temporal encoder, a planning head, and an odometry head.

The 4D spatio-temporal encoder replaces traditional mobile stack perception and prediction modules. It begins with a 3D spatial encoder that processes N omnidirectional images through a Vision Transformer (ViT) and Lift-Splat-Shoot to convert 2D image features into 3D voxel features. This 3D encoder is trained using self-supervised learning via 3D volumetric differentiable neural rendering. The 4D spatio-temporal encoder then builds upon the 3D encoder, taking past voxel features and future timestamps as input to predict future voxel features through ResNet and DiT modules, providing current and future environmental representations for planning and odometry.

The planning head, based on pre-trained 4D features, robot speed, and task information, generates executable trajectories using Transformer-based flow matching. To prevent collisions, the planning head incorporates a masked ESDF loss (Euclidean Signed Distance Field). This loss calculates the ESDF of a 3D occupancy map and applies a 2D ground truth trajectory mask, significantly reducing collision rates. Experiments demonstrate its superior performance in collision rate and overall score on out-of-distribution (OOD) datasets compared to other methods.

The odometry head predicts the robot’s relative pose using current and past 4D features and additional sensor data (e.g., IMU, wheel data). It trains a Transformer model to fuse information from different sensors. Each sensor modality is processed by a specific tokenizer, combined with modality embeddings and temporal positional embeddi…

Konten dipersingkat otomatis.

🔗 Sumber: syncedreview.com