AI/ML News & Innovations Hub

Why Scientists Can't Rebuild a Polaroid Camera [César Hidalgo]

César Hidalgo has spent years trying to answer a deceptively simple question: What is knowledge, and why is it so hard to move around? We all have this intuition that knowledge is just... information. Write it down in a book, upload it to GitHub, train an AI on it—done. But César argues that's completely wrong. Knowledge isn't a thing you can copy and paste. It's more like a living organism that needs the right environment, the right people, and constant exercise to survive. Guest: César Hidalgo, Director of the Center for Collective Learning The Big Ideas 1. Knowledge Follows Laws (Like Physics) Just as temperature and gravity follow predictable rules, so does knowledge. César outlines three laws: - Time: How knowledge grows (fast at first, then it plateaus) - Space: How knowledge spreads (it's way harder than you think) - Value: How we can measure a country's "knowledge potential" 2. You Can't Download Expertise The most memorable stories in this conversation prove that knowledge is embodied—it lives in people, teams, and organizations, not in manuals. 3. Why Big Companies Fail to Adapt César explains "architectural innovation"—the idea that small changes (like shipping books directly to customers) can require a completely different organizational structure. 4. The "Infinite Alphabet" of Economies Every skill, every industry, every capability is like a letter in an alphabet. César's research shows you can actually predict which countries will grow by counting their "letter…

TiDAR: Think in Diffusion, Talk in Autoregression (Paper Analysis)

Paper: https://arxiv.org/abs/2511.08923 Abstract: Diffusion language models hold the promise of fast parallel generation, while autoregressive (AR) models typically excel in quality due to their causal structure aligning naturally with language modeling. This raises a fundamental question: can we achieve a synergy with high throughput, higher GPU utilization, and AR level quality? Existing methods fail to effectively balance these two aspects, either prioritizing AR using a weaker model for sequential drafting (speculative decoding), leading to lower drafting efficiency, or using some form of left-to-right (AR-like) decoding logic for diffusion, which still suffers from quality degradation and forfeits its potential parallelizability. We introduce TiDAR, a sequence-level hybrid architecture that drafts tokens (Thinking) in Diffusion and samples final outputs (Talking) AutoRegressively - all within a single forward pass using specially designed structured attention masks. This design exploits the free GPU compute density, achieving a strong balance between drafting and verification capacity. Moreover, TiDAR is designed to be serving-friendly (low overhead) as a standalone model. We extensively evaluate TiDAR against AR models, speculative decoding, and diffusion variants across generative and likelihood tasks at 1.5B and 8B scales. Thanks to the parallel drafting and sampling as well as exact KV cache support, TiDAR outperforms speculative decoding in measured throughput and…

PhD Bodybuilder Predicts The Future of AI (97% Certain) [Dr. Mike Israetel]

This is a lively, no-holds-barred debate about whether AI can truly be intelligent, conscious, or understand anything at all — and what happens when (or if) machines become smarter than us. Dr. Mike Israetel is a sports scientist, entrepreneur, and co-founder of RP Strength (a fitness company). He describes himself as a "dilettante" in AI but brings a fascinating outsider's perspective. Jared Feather (IFBB Pro bodybuilder and exercise physiologist) The Big Questions: 1. When is superintelligence coming? 2. Does AI actually understand anything? 3. The Simulation Debate (The Spiciest Part) Tim says a simulation of fire doesn't get hot. They go back and forth on whether you could upload your mind to a computer — Mike says yes, Tim says absolutely not. 4. Will AI kill us all? (The Doomer Debate) Mike thinks the "AI will exterminate humanity" crowd has it backwards. His argument: any system smart enough to wage war is smart enough to realize cooperation is the winning strategy. Super-intelligent AI would want to *study* us, not destroy us. He uses the raccoon analogy to explain what agency really means. 5. What happens to human jobs and purpose? 6. Do we need suffering? In a surprisingly emotional moment, Tim asks if suffering gives life meaning. Mike's answer? "Fuck no. Desperately" Mikes channel: https://www.youtube.com/channel/UCfQgsKhHjSyRLOp9mnffqVg RESCRIPT INTERACTIVE PLAYER: https://app.rescript.info/public/share/GVMUXHCqctPkXH8WcYtufFG7FQcdJew_RL_MLgMKU1U --- TIMESTAMPS:…

There Is No Leaderboard for Safety — Andrew Gordon & Nora Petrova

People are using AI for mental health advice and life decisions, but there's no oversight and no safety ratings. We grade models on speed and smarts... but not on whether they're safe to use. Why isn't that just as important? Featuring Andrew Gordon and Nora Petrova from Prolific, discussing AI evaluation, benchmarks, and why human preference matters. 🎙️ Full episode: https://youtu.be/rqiC9a2z8Io #AIShorts #AISafety #MachineLearning

That’s a Wrap: MongoDB’s 2025 in Review & 2026 Predictions

It’s nearly the end of the year—again! That means it’s time for an end-of-year blog post that expresses disbelief at the passage of time. Which, as the saying goes, flies when you’re having fun. And definitely when you’re as busy as MongoDB was in 2025. It was a big year for the company—and more importantly, for the tens of thousands of customers and millions of developers who rely on MongoDB’s modern data platform for their most mission-critical workloads. At MongoDB, everything we do starts with our obsession with customers and their needs, and if there’s a theme to MongoDB’s 2025, it was (and will continue to be) enabling customer innovation and helping them succeed in the AI era. So here are a few highlights of how MongoDB acted on behalf of customers in 2025. From the acquisition of Voyage AI to customer success across industries, a lot happened in 2025. Let’s go!* *Read to the end for 2026 thoughts. 2025: The (MongoDB) year that was Voyage AI, modernization, and search In February, MongoDB announced the acquisition of Voyage AI, a pioneer in embedding and reranking models, to enhance the accuracy of AI applications. Integrating Voyage AI's advanced retrieval technology with MongoDB’s modern, AI-ready data platform addresses a critical challenge: LLM model hallucinations caused by a lack of context. By improving retrieval accuracy for specialized domains like finance and law, the integration enables businesses to deploy AI for mission-critical use cases. To learn more,…

The Shape of AI: Jaggedness, Bottlenecks and Salients

And why Nano Banana Pro is such a big deal

Cloud native explained: How to build scalable, resilient applications

What is cloud native? Cloud native defined The term “cloud-native computing” encompasses the modern approach to building and running software applications that exploit the flexibility, scalability, and resilience of cloud computing. The phrase is a catch-all that encompasses not just the specific architecture choices and environments used to build applications for the public cloud, but also the software engineering techniques and philosophies used by cloud developers. The Cloud Native Computing Foundation (CNCF) is an open source organization that hosts many important cloud-related projects and helps set the tone for the world of cloud development. The CNCF offers its own definition of cloud native: Cloud native practices empower organizations to develop, build, and deploy workloads in computing environments (public, private, hybrid cloud) to meet their organizational needs at scale in a programmatic and repeatable manner. It is characterized by loosely coupled systems that interoperate in a manner that is secure, resilient, manageable, sustainable, and observable. Cloud native technologies and architectures typically consist of some combination of containers, service meshes, multi-tenancy, microservices, immutable infrastructure, serverless, and declarative APIs — this list is not exhaustive. This definition is a good start, but as cloud infrastructure becomes ubiquitous, the cloud native world is beginning to spread behind the core of this definition. We’ll explore that ev…

Token-count-based Batching: Faster, Cheaper Embedding Inference for Queries

Embedding model inference often struggles with efficiency when serving large volumes of short requests—a common pattern in search, retrieval, and recommendation systems. At Voyage AI by MongoDB, we call these short requests queries, and other requests are called documents. Queries typically must be served with very low latency (typically 100–300 ms). Queries are typically short, and their token-length distribution is highly skewed. As a result, query inference tends to be memory-bound rather than compute-bound. Query traffic is pretty spiky, so autoscaling is too slow. In sum, serving many short requests sequentially is highly inefficient. In this blog post, we explore how batching can be used to serve queries more efficiently. We first discuss padding removal in modern inference engines, a key technique that enables effective batching. We then present practical strategies for forming batches and selecting an appropriate batch size. Finally, we walk through the implementation details and share the resulting performance improvements: a 50% reduction in GPU inference latency—despite using 3X fewer GPUs. Padding removal makes effective batching possible Given the patterns of query traffic, one straightforward idea is: can we batch them to improve inference efficiency? Padding removal, supported in inference engines like vLLM and SGLang, makes efficient batching possible. Most inference engines accept requests in the form (B, S), where B is the sequence number in the batch, and…

HELM Arabic

As part of our efforts to better understand the multilingual capabilities of large language models (LLMs), we present HELM Arabic, a leaderboard for transparent and reproducible evaluation of LLMs on Arabic language benchmarks. This leaderboard was produced in collaboration with Arabic.AI.

2026, the year we shine

Vukosi Marivate is a Professor of Computer Science and the ABSA UP Chair of Data Science at the University of Pretoria, South Africa🇿🇦. Vukosi leads the African Institute for Data Science and Artificial Intelligence (AfriDSAI). Additionally, he co-founded both the Deep Learning Indaba, co-founder of Lelapa AI, an African startup focused on AI for Africans […] The post 2026, the year we shine appeared first on Deep Learning Indaba .

Rethinking Pre-Training for Agentic AI with Aakanksha Chowdhery - #759

Today, we're joined by Aakanksha Chowdhery, member of technical staff at Reflection, to explore the fundamental shifts required to build true agentic AI. While the industry has largely focused on post-training techniques to improve reasoning, Aakanksha draws on her experience leading pre-training efforts for Google’s PaLM and early Gemini models to argue that pre-training itself must be rethought to move beyond static benchmarks. We explore the limitations of next-token prediction for multi-step workflows and examine how attention mechanisms, loss objectives, and training data must evolve to support long-form reasoning and planning. Aakanksha shares insights on the difference between context retrieval and actual reasoning, the importance of "trajectory" training data, and why scaling remains essential for discovering emergent agentic capabilities like error recovery and dynamic tool learning. The complete show notes for this episode can be found at https://twimlai.com/go/759.

Django tutorial: Get started with Django 6

Django is a one-size-fits-all Python web framework that was inspired by Ruby on Rails and uses many of the same metaphors to make web development fast and easy. Fully loaded and flexible, Django has become one of Python’s most widely used web frameworks. Now in version 6.0, Django includes virtually everything you need to build a web application of any size, and its popularity makes it easy to find examples and help for various scenarios. Plus, Django provides tools to allow your application to evolve and add features gracefully, and to migrate its data schema if there is one. Django also has a reputation for being complex, with many components and a good deal of “under the hood” configuration required. In truth, you can use Django to get a simple Python application up and running in relatively short order, then expand its functionality as needed. This article guides you through creating a basic application using Django 6.0. We’ll also touch on the most crucial features for web developers in the Django 6 release . What version of Python do I need? To install Django 6.0, you will need Python 3.12 or better. Ideally, you should use the most recent Python version that supports everything you want to do with your Django project, but in some cases, it may not be possible to update. If you’re stuck with an earlier version of Python, you may be able to use Django 5. Consult Django’s Python version table to find out which versions you can use. Installing Django Assuming you have Pyt…

NTT Data appoints Cristiano Rios as Director of Strategy & Operations

NTT Data has strengthened its leadership team in Brazil with Cristiano Rios, who will assume the position of Director of Strategy & Operations. Cristiano Rios has more than 20 years of experience in business consulting, specializing in strategy and operations in sectors including consumer & retail, healthcare, life sciences, and industrial.

From Python3.8 to Python3.10: Our Journey Through a Memory Leak

Image generated with ChatGPT (OpenAI), 2025. Intro When working with Python, memory management often feels like a solved problem. The garbage collector quietly does its job, and unlike C or C++, we rarely think about malloc or free. This doesn’t mean that there are no memory leaks in Python. Reference cycles, unreleased resources like connection pooling, global caches, etc can slowly inflate your process’s memory footprint. You might not notice it at first, until your worker starts OOM-ing, latency creeps up, or container restarts become mysteriously frequent. In this post, we’ll share the story of a real-world memory leak we encountered during a Python upgrade — how we discovered it, the tools and techniques we used to investigate, and the lessons we learned. What happened after upgrading to Python 3.10? Back in the summer of 2024, we had an initiative at Lyft to upgrade all of our Python services from v3.8 to 3.10 as v3.8 was scheduled to be EoL by the end of 2024. You can find more details on how our awesome Backend Foundations team at Lyft does Python upgrade across hundreds of repos at scale here . The upgrade involved two phases: the first phase was to upgrade all the dependencies to be Python 3.10 compatible, and the second phase was to upgrade the services to Python 3.10. The dependency upgrades went smoothly for all services and then the phase to upgrade all services to Python 3.10 rolled out. While all services were running Python 3.10 smoothly, there was one servi…

Titans: Learning to Memorize at Test Time (Paper Analysis)

Paper: https://arxiv.org/abs/2501.00663 Abstract: Over more than a decade there has been an extensive research effort on how to effectively utilize recurrent models and attention. While recurrent models aim to compress the data into a fixed-size memory (called hidden state), attention allows attending to the entire context window, capturing the direct dependencies of all tokens. This more accurate modeling of dependencies, however, comes with a quadratic cost, limiting the model to a fixed-length context. We present a new neural long-term memory module that learns to memorize historical context and helps attention to attend to the current context while utilizing long past information. We show that this neural memory has the advantage of fast parallelizable training while maintaining a fast inference. From a memory perspective, we argue that attention due to its limited context but accurate dependency modeling performs as a short-term memory, while neural memory due to its ability to memorize the data, acts as a long-term, more persistent, memory. Based on these two modules, we introduce a new family of architectures, called Titans, and present three variants to address how one can effectively incorporate memory into this architecture. Our experimental results on language modeling, common-sense reasoning, genomics, and time series tasks show that Titans are more effective than Transformers and recent modern linear recurrent models. They further can effectively scale to larger…

2025 Year in Review

An eventful year of progress in health and career, while making time for travel and reflection.

Why Vision Language Models Ignore What They See with Munawar Hayat - #758

In this episode, we’re joined by Munawar Hayat, researcher at Qualcomm AI Research, to discuss a series of papers presented at NeurIPS 2025 focusing on multimodal and generative AI. We dive into the persistent challenge of object hallucination in Vision-Language Models (VLMs), why models often discard visual information in favor of pre-trained language priors, and how his team used attention-guided alignment to enforce better visual grounding. We also explore a novel approach to generalized contrastive learning designed to solve complex, composed retrieval tasks—such as searching via combined text and image queries—without increasing inference costs. Finally, we cover the difficulties generative models face when rendering multiple human subjects, and the new "MultiHuman Testbench" his team created to measure and mitigate issues like identity leakage and attribute blending. Throughout the discussion, we examine how these innovations align with the need for efficient, on-device AI deployment. The complete show notes for this episode can be found at https://twimlai.com/go/758.

30% Faster Multimodal AI Training with Ray and Disaggregated Hybrid Parallelism

AI-powered assistants for scientific discovery

Andreas Geiger receives ERC Consolidator Grant

From DeepSeek V3 to V3.2: Architecture, Sparse Attention, and RL Updates

Understanding How DeepSeek's Flagship Open-Weight Models Evolved

The complete guide to Node.js frameworks

Node.js is one of the most popular server-side platforms, especially for web applications. It gives you non-blocking JavaScript without a browser, plus an enormous ecosystem. That ecosystem is one of Node’s chief strengths, making it a go-to option for server development. This article is a quick tour of the most popular web frameworks for server development on Node.js . We’ll look at minimalist tools like Express.js, batteries-included frameworks like Nest.js, and full-stack frameworks like Next.js. You’ll get an overview of the frameworks and a taste of what it’s like to write a simple server application in each one. Minimalist web frameworks When it comes to Node web frameworks, minimalist doesn’t mean limited. Instead, these frameworks provide the essential features required to do the job for which they are intended. The frameworks in this list also tend to be highly extensible, so you can customize them as needed. With minimalist frameworks, pluggable extensibility is the name of the game. Express.js At over 47 million weekly downloads on npm, Express is one of the most-installed software packages of all time—and for good reason. Express gives you basic web endpoint routing and request-and-response handling inside an extensible framework that is easy to understand. Most other frameworks in this category have adopted the basic style of describing a route from Express. This framework is the obvious choice when you simply need to create some routes for HTTP, and you don’t m…

Stop Saying Boredom is Good for Kids

Chronic boredom causes stress, disengagement, and poor well-being in adults. So why do we glorify it for children?

Scaling Agentic Inference Across Heterogeneous Compute with Zain Asgar - #757

In this episode, Zain Asgar, co-founder and CEO of Gimlet Labs, joins us to discuss the heterogeneous AI inference across diverse hardware. Zain argues that the current industry standard of running all AI workloads on high-end GPUs is unsustainable for agents, which consume significantly more tokens than traditional LLM applications. We explore Gimlet’s approach to heterogeneous inference, which involves disaggregating workloads across a mix of hardware—from H100s to older GPUs and CPUs—to optimize unit economics without sacrificing performance. We dive into their "three-layer cake" architecture: workload disaggregation, a compilation layer that maps models to specific hardware targets, and a novel system that uses LLMs to autonomously rewrite and optimize compute kernels. Finally, we discuss the complexities of networking in heterogeneous environments, the trade-offs between numerical precision and application accuracy, and the future of hardware-aware scheduling. The complete show notes for this episode can be found at https://twimlai.com/go/757.

AI Company Safety Practices Fall Short of Public Commitments and Show Structural Weaknesses, as Top Performers Widen the Gap

But in a win for transparency, five leading companies participated in the scorecard's survey for the first time, providing critical new information to the public.

Vector researchers advance AI frontiers with 80 papers at NeurIPS 2025

Researchers from Vector’s vibrant community are presenting groundbreaking work across the full spectrum of artificial intelligence at this year’s Conference on Neural Information Processing Systems (NeurIPS), taking place December 2-7 […] The post Vector researchers advance AI frontiers with 80 papers at NeurIPS 2025 appeared first on Vector Institute for Artificial Intelligence .

Peers Consulting + Technology partners with global player Andersen Consulting

Brazil-based Peers Consulting + Technology has become a collaborating firm of Andersen Consulting. Founded in 2012, Peers Consulting + Technology (Peers) is one of the fastest growing business and technology consultancies in the Brazilian market, recognized by the Financial Times for three consecutive years.

No, AI hasn’t run out of data

AI models’ relationship with our data is getting more dynamic, contextual and private—and the stakes are high The Claim Earlier this year, Elon Musk claimed that ‘all human data for AI training has been exhausted’. Ilya Sutskever, a co-founder of OpenAI, has said the world has reached ‘peak data’. A recent episode of the BBC’s […] The post No, AI hasn’t run out of data appeared first on OpenMined .

How do you identify and reduce artifacts in Neural Vocoders?

I’m evaluating a few neural vocoders (HiFi-GAN, Vocos, etc.) and I’m seeing the same type of artifact across all of them. I’m not sure exactly what it is, and I’m looking for help identifying it and figuring out how to reduce or remove it. Overall, the models can generate very good audio. In some cases the output is almost indistinguishable from the ground truth. But in other cases, there’s a consistent artifact that’s clearly audible. I’m not an audio expert, so I tried to inspect the signals visually. I plotted both the ground truth (GT) and the predicted audio (PRED) as: waveform (time domain) spectrogram (frequency domain) For the GT audio https://drive.google.com/file/d/1u_p9j29khFrnuqVoHxt8dQ9RgTcjyhpc/view?usp=sharing : The waveform looks “healthy” or “full”. The spectrogram also looks clean and well-structured, and the audio sounds perfect. For the PRED audio https://drive.google.com/file/d/14dsXOZR7JmOwnmNG3kFh6p5Cr7nagrBP/view?usp=sharing : The waveform looks much “thinner” than the GT, like it’s missing energy or detail. In the spectrogram, there are visible “smudges” or “stains” that line up with the artifacts I’m hearing. My intuition is that I might need a better loss function that explicitly penalizes these types of artifacts so the generator learns to avoid them. In the model used to generate the PRED audio, I already implemented a Multi-Resolution STFT loss, which helped to some extent, but the artifacts are still present. This seems to happen with multiple…

SEA-LION Inspires Community to Build AI Solutions That Solve Real-World Challenges in Southeast Asia

AI Singapore (AISG), together with its partners and collaborators, recently hosted the inaugural SEA-LION Summit: Powering Southeast Asia’s AI Future, a key platform demonstrating how SEA-LION is activating the...

What is devops? Bringing dev and ops together to build better software

A portmanteau of “development” and “operations,” devops emerged as a way of bringing together two previously separate groups responsible for the building and deploying of software. In the old world, developers (devs) typically wrote code before throwing it over to the system administrators (operations, or ops) to deploy and integrate that code. But as the industry shifted towards agile development and cloud-native computing , many organizations reoriented around modern, cloud-native practices in the pursuit of faster, better releases. This required a new way to perform these key functions in a more streamlined, efficient, and cohesive way, one where the old frustrations of disconnected dev and ops functions would be eliminated. With two groups working together, developers can rapidly roll out small code enhancements via continuous integration and delivery rather than spending years on “big bang” product releases. Devops was born at cloud-native companies like Facebook, Netflix, Spotify, and Amazon; but it’s become one of the defining technology industry trends of the past decade, primarily because it bridges so many of the changes that have shaped modern software development. As agile development and cloud-native computing have become ubiquitous, devops has enabled the entire industry to speed up its software development cycles. Thus, devops has now thoroughly infiltrated the enterprise, especially in organizations that rely on software to run their business, such as banks,…

Summer Fellowship 2025 Wrap Up - What Did Our Fellows Work On? | GovAI Blog

We're proud to highlight what our 2025 Summer Fellows have been working on.

The overthinking problem in AI

Reasoning models can generate seven to 10 times as many tokens as necessary on simple tasks, creating unsustainable costs at scale. Amazon's vision for metacognitive AI could fundamentally shift how models allocate computational resources.

Ray Serve LLM on Anyscale: APIs for Wide-EP and Disaggregated Serving with vLLM

New study reveals AI’s $100B economic impact across Canada, with Ontario leading the charge

TORONTO, [November 25, 2025] – Ontario is the engine of Canada’s artificial intelligence (AI) economy. AI-related jobs here in the province contributed between $42 billion and $52 billion over the […] The post New study reveals AI’s $100B economic impact across Canada, with Ontario leading the charge appeared first on Vector Institute for Artificial Intelligence .

Alliott Global Alliance expands network with Guatemala-based Palomo Abogados

Alliott Global Alliance has expanded its network in South America with the addition of Palomo Abogados, a Guatemala-based law firm. Founded in 1978, Palomo Abogados provides legal solutions to domestic and multinational corporations, regional businesses, and high-profile clients across Latin America.

Product Evals in Three Simple Steps

Label some data, align LLM-evaluators, and run the eval harness with each change.

Where did it all go wrong? A hierarchical look into multi-agent error attribution

Error attribution in Large Language Model (LLM) multi-agent systems presents a significant challenge in debugging and improving collaborative AI systems. Current approaches to pinpointing agent and step level failures in multi-agent interaction traces—whether using all-at-once evaluation, step-by-step analysis, or binary search—fall short when analyzing complex patterns, struggling with both accuracy and consistency. We present ECHO (Error attribution through Contextual Hierarchy and Objective consensus analysis), a novel algorithm that combines hierarchical context representation, objective analysis-based evaluation, and consensus voting to improve error attribution accuracy. Our approach leverages a positional-based leveling of contextual understanding while maintaining objective evaluation criteria, ultimately reaching conclusions through a consensus mechanism. Experimental results demonstrate that ECHO outperforms existing methods across various multi-agent interaction scenarios, showing particular strength in cases involving subtle reasoning errors and complex interdependencies. Our findings suggest that leveraging these concepts of structured, hierarchical context representation combined with consensus-based objective decision-making, provides a more robust framework for error attribution in multi-agent systems.

Proactive Agents for the Web with Devi Parikh - #756

Today, we're joined by Devi Parikh, co-founder and co-CEO of Yutori, to discuss browser use models and a future where we interact with the web through proactive, autonomous agents. We explore the technical challenges of creating reliable web agents, the advantages of visually-grounded models that operate on screenshots rather than the browser’s more brittle document object model, or DOM, and why this counterintuitive choice has proven far more robust and generalizable for handling complex web interfaces. Devi also shares insights into Yutori’s training pipeline, which has evolved from supervised fine-tuning to include rejection sampling and reinforcement learning. Finally, we discuss how Yutori’s “Scouts” agents orchestrate multiple tools and sub-agents to handle complex queries, the importance of background, "ambient" operation for these systems, and what the path looks like from simple monitoring to full task automation on the web. The complete show notes for this episode can be found at https://twimlai.com/go/756.

LyftLearn Evolution: Rethinking ML Platform Architecture

Written by Yaroslav Yatsiuk At Lyft, machine learning (ML) is the engine behind our most critical business functions — from dispatch and pricing optimization to fraud detection and support automation. Our ML infrastructure serves thousands of production models making hundreds of millions of real-time predictions per day, supported by thousands of daily training jobs that keep ML models fresh and accurate. As our scale grew, we faced a classic engineering challenge: the very complexity that powered our platform was becoming a bottleneck to its future growth. We needed to answer a fundamental question: How could we evolve our platform to accelerate innovation for our users while simplifying its underlying architecture? This post explores how we rethought LyftLearn’s architecture to solve this problem. We’ll walk through our transition from a fully Kubernetes-based system to a hybrid platform, combining the simplicity of managed compute on AWS SageMaker for offline workloads with the flexibility of Kubernetes for online model serving. Afterwards, we’ll share the key technical decisions and trade-offs that made this evolution possible. LyftLearn Overview LyftLearn is Lyft’s end-to-end machine learning platform, managing the complete ML lifecycle from model development to production serving. Built to support hundreds of data scientists and ML engineers, it handles the full spectrum of ML workloads at scale. The platform is composed of three integrated products: Figure 1: LyftLear…

Three Years from GPT-3 to Gemini 3

From chatbots to agents

Is the strong duality of the hard-margin SVM really trivially satisfied all the time?

It is widely known that if you were to calculate the maximizer of the dual SVM program (denote as $\alpha^*$ ), then the primal minimizer of the hard-margin SVM program, \begin{aligned}&{\underset {{w} ,\;b }{\operatorname {minimize} }}&&\|{w} \|_{2}^{2}\\&{\text{s.t.}}&&y_{i}({w} ^{\top }{x} _{i}-b)\geq 1\quad \forall i\in \{1,\dots ,N\}\end{aligned} can be found through the KKT conditions, in which case, $$w^*= \sum_{i = 1}^N \alpha^*_i y_i x_i$$ In particular, those $x_i$ corresponding to non-zero $\alpha^*_i$ are the famous support vectors. However, a technical detail that people seem to ignore is that this only holds if strong duality also holds. I struggled with verifying strong duality, so I searched widely online. Essentially, online reference argue that Slater's condition guarantees strong duality always hold for hard-margin SVM. Slater condition in this case can be written as either: There exists a pair $(w,b)$ that satisfies $y_{i}({w} ^{\top }{x} _{i}-b)> 1$ for all $i = 1, \ldots, N$ , or There exists a pair $(w,b)$ that satisfies $y_{i}({w} ^{\top }{x} _{i}-b)\geq 1$ , for all $i = 1, \ldots, N$ The latter is because the constraint functions are affine. If Slater holds, then strong duality holds and we can make the claim $$w^*= \sum_{i = 1}^N \alpha^* y_i x_i.$$ If it doesn't, then this relationship does not hold. I do not think Slater always holds despite many answers in the contrary online (virtually all of them). I suspect they are making a mistake. I think…

Outstanding students at the Federal Competition for Artificial Intelligence

Diagnosing strokes faster, opening the black box of language models, or determining blood sugar levels painlessly – winning teams provide forward-looking impetus with AI projects

When smart AI gets too smart: Key insights from Vector’s 2025 ML Security & Privacy Workshop

Vector Institute’s 2025 Machine Learning Security & Privacy Workshop revealed critical AI safety breakthroughs and concerning vulnerabilities in current machine learning (ML) security methods. This comprehensive analysis covers the latest […] The post When smart AI gets too smart: Key insights from Vector’s 2025 ML Security & Privacy Workshop appeared first on Vector Institute for Artificial Intelligence .

Altman Solon deepens Brazilian presence with new office in São Paulo

Altman Solon, a global strategy consultancy specialising in telecommunications, media, and technology (TMT), has launched a new office in São Paulo – its second in Latin America. The move in the Brazilian market formalizes the firm’s longstanding work with clients in the country – Altman Solon has been advising local companies and investors for over a decade.

Mixed Methods Scenario Development for Human-Vehicle Interaction Research: A Case Study on Winter Driving

Mixed Methods Scenario Development for Human-Vehicle Interaction Research: A Case Study on Winter Driving robyn.cherinka… Wed, 11/12/2025 - 14:59 Scenarios provide a fundamental link between driving simulators and real-world conditions, shaping the extent to which the findings of a user study can be applied to public roads. However, compared to other aspects of study design, scenario development in human–vehicle interaction research tends to receive less deliberate attention. To encourage more methodical scenario generation, this work introduces a mixed methods approach for extracting representative scenarios from an integration of three real-world data sources: aggregated crash statistics, interviews with experienced drivers, and naturalistic driving data. Through a case study on winter driving, we outline the derivation of a nighttime, two-lane road scenario from these data sources and conduct an initial driving simulator pilot study to assess its realism. We hope that this demonstration of scenario generation from quantitative and qualitative data inspires researchers to consider more rigorous methods for scenario design in future work. Read More Image Oct 8, 2025 Human Interactive Driving 1 Minute Read

Robot Learning from Any Images

Robot Learning from Any Images robyn.cherinka… Wed, 11/12/2025 - 14:54 We introduce RoLA, a framework that transforms any in-the-wild image into an interactive, physics-enabled robotic environment. Unlike previous methods, RoLA operates directly on a single image without requiring additional hardware or digital assets. Our framework democratizes robotic data generation by producing massive visuomotor robotic demonstrations within minutes from a wide range of image sources, including camera captures, robotic datasets, and Internet images. At its core, our approach combines a novel method for single-view physical scene recovery with an efficient visual blending strategy for photorealistic data collection. We demonstrate RoLA's versatility across applications like scalable robotic data generation and augmentation, robot learning from Internet images, and single-image real-to-sim-to-real systems for manipulators and humanoids. Video results are available at this https URL . Read More Image Sep 26, 2025 Robotics 1 Minute Read

From Dashboards to Dialogue: Evaluating a Conversational AI Coach for Performance Driving Skill Development

From Dashboards to Dialogue: Evaluating a Conversational AI Coach for Performance Driving Skill Development robyn.cherinka… Wed, 11/12/2025 - 14:50 Learning in domains involving complex motor skills, such as performance driving, often requires feedback that is timely, personalized, and actionable. Yet many drivers rely on video and telemetry data to review their performance without guidance. We explore how conversational AI can support post-drive reflection by integrating LLM-generated coaching into an interactive review interface. In an exploratory within-subjects simulator study (n=16), participants completed laps under two conditions: one with video and data visualizations alone, and another with the same tools augmented with a conversational interface that provided verbal feedback after each lap. Conversational feedback supported short-term improvements in lap time, average speed, and steering control, and was rated as more useful and satisfying—though it also elicited slightly higher nervousness. These results suggest that conversational AI can make post-drive feedback more interpretable and actionable, particularly for drivers reviewing performance data in high-skill contexts like performance driving. Read More Image Oct 4, 2025 Human Interactive Driving 1 Minute Read

SAFE: Multitask Failure Detection for Vision-Language-Action Models

SAFE: Multitask Failure Detection for Vision-Language-Action Models robyn.cherinka… Wed, 11/12/2025 - 14:45 While vision-language-action models (VLAs) have shown promising robotic behaviors across a diverse set of manipulation tasks, they achieve limited success rates when deployed on novel tasks out of the box. To allow these policies to safely interact with their environments, we need a failure detector that gives a timely alert such that the robot can stop, backtrack, or ask for help. However, existing failure detectors are trained and tested only on one or a few specific tasks, while generalist VLAs require the detector to generalize and detect failures also in unseen tasks and novel environments. In this paper, we introduce the multitask failure detection problem and propose SAFE, a failure detector for generalist robot policies such as VLAs. We analyze the VLA feature space and find that VLAs have sufficient highlevel knowledge about task success and failure, which is generic across different tasks. Based on this insight, we design SAFE to learn from VLA internal features and predict a single scalar indicating the likelihood of task failure. SAFE is trained on both successful and failed rollouts, and is evaluated on unseen tasks. SAFE is compatible with different policy architectures. We test it on OpenVLA, π0, and π0-FAST in both simulated and real-world environments extensively. We compare SAFE with diverse baselines and show that SAFE achieves state-of-the-art failu…

STITCH-OPE: Trajectory Stitching with Guided Diffusion for Off-Policy Evaluation

STITCH-OPE: Trajectory Stitching with Guided Diffusion for Off-Policy Evaluation robyn.cherinka… Wed, 11/12/2025 - 14:40 Off-policy evaluation (OPE) estimates the performance of a target policy using offline data collected from a behavior policy, and is crucial in domains such as robotics or healthcare where direct interaction with the environment is costly or unsafe. Existing OPE methods are ineffective for high-dimensional, long-horizon problems, due to exponential blow-ups in variance from importance weighting or compounding errors from learned dynamics models. To address these challenges, we propose STITCH-OPE, a model-based generative framework that leverages denoising diffusion for long-horizon OPE in high-dimensional state and action spaces. Starting with a diffusion model pre-trained on the behavior data, STITCH-OPE generates synthetic trajectories from the target policy by guiding the denoising process using the score function of the target policy. STITCH-OPE proposes two technical innovations that make it advantageous for OPE: (1) prevents over-regularization by subtracting the score of the behavior policy during guidance, and (2) generates long-horizon trajectories by stitching partial trajectories together end-to-end. We provide a theoretical guarantee that under mild assumptions, these modifications result in an exponential reduction in variance versus long-horizon trajectory diffusion. Experiments on the D4RL and OpenAI Gym benchmarks show substantial improveme…

AI Orchestration for Smart Cities and the Enterprise with Robin Braun and Luke Norris - #755

Today, we're joined by Robin Braun, VP of AI business development for hybrid cloud at HPE, and Luke Norris, co-founder and CEO of Kamiwaza, to discuss how AI systems can be used to automate complex workflows and unlock value from legacy enterprise data. Robin and Luke detail high-impact use cases from HPE and Kamiwaza’s collaboration on an “Agentic Smart City” project for Vail, Colorado, including remediation and automation of website accessibility for 508 compliance, digitization and understanding of deed restrictions, and combining contextual information with camera feeds for fire detection and risk assessment. Additionally, we discuss the role of private cloud infrastructure in overcoming challenges like cost, data privacy, and compliance. Robin and Luke also share their lessons learned, including the importance of fresh data, and the value of a "mud puddle by mud puddle" approach in achieving practical AI wins. The complete show notes for this episode can be found at https://twimlai.com/go/755.