earuphoria.
All podcasts
News · Technology

UpNext AI

UpNext Labs
Rights holder? Report thisupnext.fm

Episodes

November 18, 2026 · Episode 143

OpenAI’s GPT-6.2 Astra and Sol, ArisGlobal’s Bedrock Costs, and Nvidia’s Scientific Agents

OpenAI’s GPT-6.2 Astra and Sol launch across ChatGPT Work, Codex and the API; Latent Space examines independent results and early user reactions. ArisGlobal’s NavaX on Amazon Bedrock, the 3CBench infrastructure benchmark, Calif’s Java vulnerability discovery, Simon Willison’s coding-agent sandboxes, First Person Labs’ user models and Nvidia’s SC26 scientific-computing announcements round out the episode. OpenAI introduces Astra and Sol, with new professional-work and coding capabilities, API pricing changes and preparedness assessments.Latent Space reviews Artificial Analysis results and conflicting early experiences with the new models.ArisGlobal describes how caching shared prompts changed inference spending for NavaX on...

November 17, 2026 · Episode 142

OpenAI’s NATO Cyber Tools, Researcher Access, and the U.S. AI Chip Access Rule

OpenAI’s Daybreak work with NATO’s Communications and Information Agency, ENISA, and CERT Polska leads today’s episode. We also cover ChatGPT for Academic Researchers, a Nature Computational Science commentary on AI in mathematics, Charles Xiao’s study of agent guardrails, the U.S. Commerce Department’s remote AI-chip access rule, California’s OpenAI lawsuit involving Hugging Face, OpenAI’s embedded assessments, and a fix for Simon Willison’s LLM tool. OpenAI describes how Daybreak is being used in allied cyber-defense work, including vulnerability investigations and a NATO training exercise.ChatGPT for Academic Researchers expands with a new application...

November 16, 2026 · Episode 141

OpenAI and LVMH’s Journalism Deal, Nvidia’s AI Day Seoul, and XR Blocks

OpenAI and LVMH’s Media Division are bringing ChatGPT Enterprise into publication workflows and linking ChatGPT users to journalism; Nvidia’s AI Day Seoul highlighted robotics and agents, while executives at Meta, Jack Henry & Associates and PwC discussed AI spending. Also covered: XR Blocks and Vibe Coding XR research, Meituan’s LongCat-2.5, Nvidia’s open-model ecosystem, allied AI supply chain security and Japan’s planned Defense AI Vision. OpenAI and LVMH’s Media Division sign an agreement spanning newsroom tools and journalism access through ChatGPT. [Source](https://openai.com/index/lvmh-media-division-partnership)Nvidia recaps robotics, agent demonstrations and developer training at...

November 13, 2026 · Episode 140

Amplitude’s Wave, Nvidia’s Hugging Face Deal, and Metal Sandbox

Amplitude’s Wave connects OpenAI-powered product analysis with coding agents and impact measurement; Hugging Face CEO Clément Delangue responds to Jensen Huang’s departure amid Nvidia’s pending acquisition. Also covered: Nectar’s Creator Agent, Metal Sandbox for Apple Silicon, Anthropic’s Claude compute strategy, and JioStar chief architect Vijay Seshadri on AI discovery. Amplitude’s Wave proposes product changes, hands plans to coding agents, and measures results. [Source](https://openai.com/business/customer-stories/amplitude)Delangue pledges continued support for open-source AI as Nvidia’s planned Hugging Face acquisition remains pending. [Source](https://techcrunch.com/2026/11/12/hugging-face-ceo-vows-to-keep-fighting-for-open-source-ai-after-jensen-huangs-exit/)Nectar launches Creato...

November 12, 2026 · Episode 139

Anthropic’s AI Acceleration Warning, TSMC’s OpenAI and Anthropic Investment Plan, and Pyrite’s Drug-Binding Benchmark

The Information reports heightened concern about AI’s pace, while TSMC plans investments in OpenAI and Anthropic. Also covered: OpenAI’s children’s safety research commitment, the Pyrite protein–ligand pose-ranking benchmark, Alibaba’s chip-access-related share decline, Google DeepMind’s AlphaGenome 2, and inference startups buying chips. The Information on AI acceleration and safety researchers’ concerns: https://www.theinformation.com/articles/anthropic-others-warn-ai-accelerationTSMC’s planned OpenAI and Anthropic investments: https://arstechnica.com/ai/2026/11/taiwans-chipmaking-giant-tsmc-to-invest-1b-in-anthropic-openai/OpenAI’s commitment to children’s AI safety research: https://openai.com/index/children-first-declarationPyrite’s benchmark for ranking candidate protein–ligand binding poses: https://arxiv.org/abs/2611.14511v1Alib...

November 11, 2026 · Episode 138

Microsoft’s Lazarus Backdoor, Gemini Robotics 2.5, and Mosaic’s Surgical Catheter

Microsoft researchers describe Lazarus, a hidden-trigger backdoor for AI agents, while Google DeepMind introduces Gemini Robotics 2.5 and the ASIMOV-jailbreak safety benchmark. Also covered: Mosaic Intelligence Labs’ surgical catheter, a study of enterprise analytics agent evaluations, Nvidia chip routes through Singapore, Google’s Adobe integration, and CoreWeave’s planned chip-backed financing. Microsoft’s Lazarus research examines malicious edits to model weights and the importance of provenance and agent permissions.Google DeepMind’s Gemini Robotics 2.5 pairs dexterous-task demonstrations with an open safety benchmark.Mosaic Intelligence Labs has raised $47 million while developing a semi-autonomous surgical catheter.An enterprise-agent evaluation preprint compares trace-base...

November 10, 2026 · Episode 137

CodeRabbit’s Codex Output, OpenAI’s Control Room, and Meta’s Muse Ember

CodeRabbit CEO Harjot Gill discusses Codex-assisted software delivery and CodeRabbit Secure, while OpenAI details its Control Room for stopping research workloads and Meta introduces Muse Ember for repetitive agent tasks. Also covered: DeidSharp’s clinical-text de-identification evaluation, proposed Nvidia chip export controls, security executive Lisa Thiergart’s assessment of espionage risks, and a Claude Fable 5.2 prompting test. CodeRabbit describes shipping roughly twenty times more software with the same engineering team and explains how it reviews and secures agent-generated code. Source: https://openai.com/index/coderabbit-codex-daybreakOpenAI explains how Control Room tracks research jobs, denies inference requests for stopped jobs, and...

November 9, 2026 · Episode 136

NVIDIA and Korea’s Sovereign AI Push, Anthropic’s Power Deal, and OpenAI’s Hacktron Partnership

NVIDIA and Korea’s Ministry of Science and ICT are expanding work on sovereign AI infrastructure, NVLink Fusion and Korean models, while NVIDIA names investments involving Point2 Technology and plans for Motif Technologies, RLWRLD and Upstage. Also covered: Anthropic’s reported gas-fired power deal, Yoko Hasebe’s proposed autonomous-system controls, OpenAI’s Hacktron partnership, Simon Willison’s AI pull-request workflow, and the reported Crusoe–Google data center sale. NVIDIA and Korea expand collaboration on chips, infrastructure, inference software and skills. https://blogs.nvidia.com/blog/korea-sovereign-ai-collaboration/NVIDIA announces a Point2 investment and plans further Korean startup investments. https://blogs.nvidia...

November 6, 2026 · Episode 135

OpenAI’s GPT-6 Cyber, Agricultural Agents, and XStack’s ML Migration Benchmark

OpenAI’s GPT-6 Cyber pairs vulnerability research with authorized-scope safeguards, while an npj Sustainable Agriculture review examines agricultural agents and the SecureLLM study tests compact models against cybersecurity archives. XStack measures PyTorch-to-JAX repository migration; Brian Chesky’s AI startup, Dropbox’s Tigris Data acquisition, and Claude’s Plaid-connected personal finance feature round out the episode. OpenAI reports GPT-6 Cyber benchmark gains, Firefox vulnerability findings, Linux kernel exploit work, and Daybreak Red safeguards.A review maps potential uses and deployment challenges for AI agents in crop management, agricultural research, and robotics.SecureLLM researchers test continued training and retrieval using archived...

November 5, 2026 · Episode 134

Realtime-Bench’s Agent Test, Google’s Genie 4, and OpenAI’s Codex Acquisition

Realtime-Bench tests time-sensitive AI agents, while Google DeepMind previews Genie 4; OpenAI is acquiring Pierre Computer Company for Codex and offering GPT-5.6 Cyber through Daybreak Red on Amazon Bedrock in AWS GovCloud. Also covered: an arXiv study of text-embedding compression, Cyera’s discovery of unnoticed agents, Palo Alto Networks’ reported talks with Clutch Security, and DuckDuckGo’s pitch to schools. Realtime-Bench compares agent performance with running and paused clocks; Google DeepMind previews changes to interactive Genie 4 simulations. Source: https://importai.substack.com/p/import-ai-479-spacewar-low-latencyOpenAI plans to acquire Pierre Computer Company, whose Trees and Diffs libraries are used in Codex...

November 4, 2026 · Episode 133

Ballistic Ventures’ Cybersecurity Funds, Cisco’s OpenAI Defense, and Parkinson’s Screening

Ted Schlein’s Ballistic Ventures is seeking new cybersecurity funds as CrowdStrike plans acquisitions of Fiddler AI and Dreadnode; Cisco describes its security work with OpenAI, while a Nature Aging paper introduces Quest2Dx for Parkinson’s questionnaire screening. Also covered: the DEP language-model evaluation toolkit, Anthropic’s FlashProve and Message Threads, Google’s school AI controls, and The Information’s reporting on OpenAI, Nvidia and chip sales to China. Ballistic Ventures targets $490 million across two funds; CrowdStrike plans to buy two AI security startups.Cisco and OpenAI describe network-informed security work and plans for the Daybreak Defense Network.A...

November 3, 2026 · Episode 132

Legora’s Legal AI, Weight-Space Programming, and WebHarbor’s Browser-Agent Tests

Legora CTO Jacob Lauritzen discusses confidential legal AI evaluations and Tabular Review; Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab report early weight-space programming experiments; and WebHarbor offers repeatable environments for browser-agent tests. Also covered: Anthropic’s Claude Code game-recreation rule, Simon Willison’s LLM release, OpenAI’s bug bounty, Marvell’s Lumentum deal, and Oracle’s AI infrastructure financing. Legora’s approach to evaluating legal agents without access to customer documents, and Lauritzen’s account of changing legal billing.Early experiments in reading and editing neural-network weights directly, including a reported repair to Qwen’s reasoning loops.Claude Cod...

November 2, 2026 · Episode 131

Salesforce’s Rune Deal, Open-Model Cyber Risks, and AI’s Water Footprint

Salesforce’s proposed Rune Technologies acquisition would extend Missionforce into field logistics for the U.S. military. Also covered: Nathan Lambert on open-weight models and cybersecurity, the Anthropic Claude extension incident in Cursor, a review of AI’s water footprint, and updates involving Claude watermarks, ChatGPT, Datasette Agent and empathetic chatbot prompts. Salesforce agrees to acquire AI logistics startup Rune Technologies in a reported deal worth $375 million.Nathan Lambert examines open-weight models, vulnerability discovery and the wider cyber response.A malicious version of Anthropic’s Claude extension was distributed through Cursor’s marketplace.A narrative review examines direct and indi...

October 30, 2026 · Episode 130

Thrive Holdings’ ADT Deal, Mistral Large 4, and Meta’s Style Sketch

Thrive Holdings’ ADT acquisition puts OpenAI-backed investment into home security, while Mistral Large 4 targets organization-controlled cyber defense and Meta’s Style Sketch builds outfits from Instagram photos. We also cover the Ramanujan Challenge’s AI-assisted proof evaluation, ChatGPT API credits, DeepSeek’s funding, Google’s earnings reaction, and Spawn’s Fable-powered coding agents. Thrive Holdings’ ADT deal: enterprise value of $15.3 billion and plans for a smart-home security business. https://techcrunch.com/2026/10/29/openai-backed-thrive-holdings-acquires-adt/Mistral Large 4: downloadable model weights and a cybersecurity deployment pitch. https://techcrunch.com/2026/10/29/mistral-releases-weights-for-its-trillion-parameter-ai-model-targeting-cybersecurity/Meta Style Sketch: a virtual closet and outfit suggestions using users’ clothes. https://techcrun...

October 29, 2026 · Episode 129

Anthropic’s AI Spending Reports, Microsoft’s OpenAI Revenue, and Google’s Search Ads

Anthropic’s enterprise reporting tools, Microsoft’s OpenAI revenue and Google’s AI Mode advertising expansion lead today’s episode. We also cover a latent world-model evaluation, Atomicwork’s cybersecurity hiring plans and GitHub’s Octoverse activity figures. Anthropic adds reports on team output, spending and workflow friction, alongside pooled token budgets.Microsoft reports $10 billion in quarterly revenue from commercial arrangements with OpenAI.Google expands AI Mode ads internationally as Search advertising growth slows; Sundar Pichai says AI features are contributing to query growth.Researchers use latent rollout consistency to test whether world models faithfully predict environmental dynamics.Atomicwork p...

October 28, 2026 · Episode 128

Retell’s OpenAI Voice Agents, Google’s TPU Demand, and Anthropic’s IPO Plans

Retell’s OpenAI-powered voice agents, Google TPUs reaching Oracle, Anthropic’s IPO plans and TeamGym’s tests of AI-agent coordination lead today’s episode. We also cover Google DeepMind’s Genie 4, AI mathematics and encryption planning, and Apple’s iOS 27.2 update. OpenAI’s Retell case study describes call volume, model use, voice-agent evaluations and engineering use of Codex. https://openai.com/index/retell-aiThe Information reports growing demand for Google TPUs, including Oracle’s plans to offer the chips and reported plans for OpenAI to buy them. https://www.theinformation.com/articles/google-tpu-momentThe Verge examines Anthropic’s IPO plans and its stated com...

October 27, 2026 · Episode 127

AWS SageMaker Edge Models, Greptile’s GPT-6 Astra Testing, and Nvidia’s Meta Investment

AWS outlines a SageMaker SDK v3 workflow for camera-specific edge models; Greptile describes GPT-6 Astra application testing and Codex codebase context. Research examines lifespan-aware AI safety evaluation and human-in-the-loop accuracy auditing, while Nvidia, Meta, OpenAI, Cohere and Alibaba feature in today’s headlines. • AWS’s reference architecture trains, checks, compiles and versions computer-vision models by camera. • Greptile describes browser-based testing with GPT-6 Astra, security work with Daybreak and codebase context for Codex. • Casey Williams and Jennifer Scurrell propose a Lifespan Vulnerability Matrix and external AI evaluation framework. • Researchers test confidence-guided human auditing and the costs of mod...

October 26, 2026 · Episode 126

OpenAI’s Frontier-Model Containment, Kimi K3.1 Pro, and AI Coding Reviews

OpenAI details changes to frontier-model training containment, and Moonshot’s Kimi K3.1 Pro reaches the Kimi API with model weights promised later. This episode also covers machine-learning-based zero-trust access, research on coding tools including GitHub Copilot, STEM-Xpert’s education offering, and Andy Jassy’s comments on Amazon Web Services capacity. OpenAI’s offline training environment, layered containment and high-compute red-teaming: https://openai.com/index/securing-our-frontier-model-training-environmentKimi K3.1 Pro’s API release, token pricing and planned weights release: https://simonwillison.net/2026/Oct/23/kimi-k31-pro/A systematic review of adaptive trust decisions in zero-trust security: https://doi.org/10.63158/journalisi.v8i5.1815A systemat...

October 23, 2026 · Episode 125

Amazon SageMaker Training Sessions, OpenAI’s NASK Partnership, and Nvidia Minerva

Amazon SageMaker AI training sessions give developers control over custom model-training loops, OpenAI and Poland’s NASK National Research Institute are exploring AI-safety research, and Nvidia Minerva brings infrastructure and workload monitoring together for AI factories. Also covered: a breast-lesion literature retrieval study, OpenAI funding talks involving UAE investors, Anthropic’s Agent Skills donation to the Agentic AI Foundation, and 1Password’s agent login feature. AWS explains how SageMaker AI training session APIs support custom training loops on managed GPUs. https://aws.amazon.com/blogs/machine-learning/write-your-own-training-loop-without-managing-gpus-with-amazon-sagemaker-training-sessions/OpenAI and NASK sign an agreement to explore AI safety and securi...

October 22, 2026 · Episode 124

Google Cloud’s Vera Rubin Rollout, OpenAI’s Biology Agents, and AI-Discovered PHP Flaws

Google Cloud and NVIDIA’s Vera Rubin rollout, OpenAI’s GPT-Rosalind work with PerturbAI and Scripps Research, and AWS’s Qwen3 training walkthrough on Trainium lead today’s episode. We also cover the TIDE model-evaluation paper, NVIDIA’s financial-services deployments, AI-discovered PHP flaws, Recursive’s Flint model, and Walmart’s shopping-agent controls. Google Cloud and NVIDIA plan general availability of Vera Rubin-based A5X instances next month, with European expansion planned. Source: https://blogs.nvidia.com/blog/google-cloud-vera-rubin-infrastructure/OpenAI, PerturbAI, and Scripps Research used distributed agents to identify testable leads in mouse brain-cell data. Source: https://openai.com/index/scaling...

Similar podcasts