Latest in AI

A food or drink brand can win a new retail listing and still struggle to fill it. Ingredients must arrive, packaging has to match specification, production needs a slot, and suppliers often need paying before the finished goods generate cash. Opply has raised $7.5 million to put AI agents to work across those connected operations, following what the company reports as a 15-fold increase in gross revenue over 12 months. Chalfen Ventures led the funding round, joined by Index Ventures, Anthemis…

Search strategy used to be straightforward. A customer had a question, typed it into a search engine, and brands competed to rank highest on the results page. But this model is quickly changing. A growing share of discovery now happens inside AI systems that interpret information, synthesize sources and provide an answer before a customer ever reaches a brand’s website. Generative Engine Optimization (GEO) has emerged as one response, but companies risk thinking too narrowly about what that…

Ayush Sharma is the Founder and CEO of Warp, which he launched in 2023 after experiencing the complexity of managing payroll across multiple U.S. states and countries. Raised in a small town in India, he studied computer science and physics at MIT. His master's research focused on mitigating slow-running components in distributed machine learning training, exploring how large-scale systems can operate more efficiently when individual machines fall behind. Warp joined Y Combinator's Winter 2023…

A suspicious login can leave clues in an identity platform, an endpoint tool, and a cloud account without any one system showing the whole incident. Mave is building AI agents to connect those clues where they already live. The Tel Aviv startup has announced $34 million in funding, including a $28 million Series A, to expand an agentic security operations platform that works across an enterprise’s existing tools. Red Dot Capital Partners led the Series A, with participation from Lumir Growth…

A lot of the material that I trawl at Arxiv and other sources every day is rather dry, even for a nerd's sensibilities; therefore I'm always looking for a publication that engages me emotionally, as a user, proponent, or victim of AI. Today that moment of hackle-activation was supplied by a new academic collaboration from the US, examining the extent to which agentic frontier AI models burn their bridges before taking damaging action – and faster than you can type 'STOP!' (hint: keep it handy…

Crusoe on October 22, 2026 announced a strategic partnership with Paraguayan energy and infrastructure developer Araico to build high-performance AI compute infrastructure in Paraguay powered by the Itaipú Hydroelectric Dam, with an initial deployment planned online in mid-2027 and a stated path to 106 MW of IT capacity. The company said the initiative is its first dedicated AI Cloud deployment in Latin America. Crusoe said the facility will be located adjacent to the Itaipú substation and will…

The Allen Institute for AI released OlmoEarth v1.3 on October 22, 2026, an open satellite embedding model accompanied by precomputed annual embeddings for the whole planet at 10-meter resolution and a browser-based OlmoEarth Embeddings Explorer that requires no sign-in. The release includes the OlmoEarth v1.3 Base model, precomputed global embeddings in 128- and 64-dimensional versions, the Explorer, and a technical report, all under the OlmoEarth Artifact License used for OlmoEarth v1 and…

Physical AI startup Enact Intelligence emerged from stealth on October 22, 2026, announcing $30 million in pre-seed funding to advance its force-and-touch approach to robotic dexterity. The Zurich-based company is building robots that can sense force and touch in addition to seeing their surroundings. Venture firms redalpine, Notion Capital and firstminute capital backed the round. Angel investors named in the announcement include Arthur Mensch, founder of Mistral; Yann LeCun, founder of AMI…

Agility Robotics announced on October 22, 2026 that its Digit 5 humanoid robot will standardize its onboard compute on NVIDIA IGX Thor modules, a transition that establishes NVIDIA as the exclusive provider of onboard compute for the general availability version of the robot. The new architecture will unify Digit 5's onboard compute on IGX Thor modules as Agility prepares its next-generation humanoid for broader commercial deployment. "Bringing humanoid robots to commercial scale requires AI…

Getting the new words right is only part of changing a sign in a video. You also want to keep the shot you chose. Suppose an AI editor replaces the lettering, and the preview looks convincing. Before accepting it, you still need to establish what happened to the rest of the footage. That review extends beyond spelling. As the sign turns, the words have to stay with its surface, while the scene around it remains as you intended. A good still frame tells you very little about the transition…

Diffusion transformers use transformer backbones inside diffusion-style generative processes, modeling noisy image, video, or latent tokens with scalable attention. This guide explains the mechanism, trade-offs, evaluation, and controls that matter in practice.

A threat advisory can reach a small utility in seconds. Deciding whether it matters—and what to do about it—can take far longer. Cyware is targeting that gap with Last Mile Agents, a new set of AI agents designed to turn shared cybersecurity intelligence into daily priorities for smaller organizations. Announced on October 22, the agents are aimed at members of threat-sharing networks that receive advisories from community or state hubs but lack a dedicated threat intelligence team. Rather than…

From Intelligence to Authority Artificial intelligence is moving beyond answering questions and generating content. Increasingly, AI systems can interpret instructions, formulate plans, select tools, and execute tasks without continuous human direction. That distinction matters. An AI system that recommends action is different from one authorized to carry it out. The difference is not necessarily intelligence or sophistication. It is authority. Autonomous systems are not limited to frontier…

IBM released Granite Speech 5.0 2B, an Apache 2.0 speech recognition and translation model. It ranks #2 on the Open ASR Leaderboard with 5.68% average English WER, alongside updated Turbo CTC and 4.1 2B Plus models.

The post IBM AI Releases Granite Speech 5.0 2B for Multilingual Transcription and Translation appeared first on MarkTechPost.

Perplexity and NVIDIA introduced JOLT, a post-training method that trains one model as both teacher and student. On MATH it matched GRPO’s accuracy with 6.5x fewer processed training tokens.

The post Perplexity and NVIDIA Researchers Introduce JOLT: a Joint Teacher-Student Reinforcement Learning Method appeared first on MarkTechPost.

Explore a comprehensive coding guide benchmarking Cactus Compute's 16.9 MB speech model against Whisper base. Learn about word error rates, CPU speed, streaming latency, and speech-to-tool integration.

The post A Coding Guide to Cactus Whistle: Does a 16.9 MB Speech Model Hold Up Against Whisper Base? Word Error Rate, Speed, Streaming, Timestamps, Keywords, and Speech-to-Tool Calls on a CPU appeared first on MarkTechPost.

Discover NVIDIA Nsight Python, an open-source interface that allows developers to profile GPU kernels, analyze hardware metrics, evaluate cache and clock controls, and measure kernel fusion directly from Python code.

The post NVIDIA Nsight Python Tutorial: How to Profile GPU Kernels appeared first on MarkTechPost.

Recursive has released Flint, an Apache 2.0 local coding model with 40B total and 2.5B active parameters. It reports 74.2% on SWE-bench Verified and 125+ tokens per second on an M4 Pro MacBook Pro.

The post Recursive Releases Flint, an Open-Weight 40B MoE Local Coding Model for Agents with 2.5B Active Params appeared first on MarkTechPost.

Vercel released durable-harness, an open-source package that logs every model and tool call so Pi, OpenCode, Claude Code and Codex sessions resume after crashes, timeouts and redeploys on Vercel, Postgres or a laptop.

The post Vercel Open-Sources durable-harness for Crash-Proof Coding Agents appeared first on MarkTechPost.

Tesla released its third-quarter 2026 financial results on October 21, 2026, reporting total revenue of $28.0 billion and GAAP operating income of $718 million in a Form 8-K filed with the SEC. The results cover the quarter ended September 30, 2026, and were published as the company's Third Quarter 2026 Update, attached to the filing as Exhibit 99.1. Third Quarter Financial Results Total revenues were $28,046 million, compared with $28,095 million in the same quarter of 2025, a result Tesla…