EVERY STORY. ONE PLACE.
Browse all 295 articles by category.
AI topic guides
- AI agents: a guide to tools, workflows and control
- AI models: understand LLMs, retrieval and evaluation
- AI coding: assistants, agents and application builders
- Creative AI: images, video, voice and production workflows
- AI at work: research, documents, meetings and useful automation
- AI in India: languages, founders and practical adoption
- AI in the USA: models, builders and infrastructure
- AI in the UK: research, founders and practical adoption
- AI in China: model releases, open weights and language
- AI in Japan: language, research and industry
- Responsible AI: evidence, privacy, oversight and accountability
- AI infrastructure: inference, local models, latency and cost
About the publication
Authors
News
- Gemini Tool Retrieval enters preview: smaller prompts, on-demand tools and the cost tradeoff
- AssemblyAI STT 3.7 arrives with three models: prices, speaker labels and migration dates
- Tasks in Gemini goes free in India: what the rollout includes and what stays U.S.-only
- Qwen4-Max leaves preview: API access, regional pricing and the limits to check
- Gemini 4 Hydrogen explained: pricing, free access and what changes in Google Search
- Suno Bridge brings AI audio into Ableton Live: access, credits and the upload to know
- ChatGPT Wallet launches: who can use it and what purchase approval means
- DeepSeek opens its app toolkit: files, links and voice for custom chatbots
- Gemini Decisions API: Google adds fast scoring for choices, routing and checks
- OpenAI brings businesses into ChatGPT: Business Hub, agents and what changes
- Meta’s Muse Ember targets high-volume agents, with a long-context pricing catch
- Claude Fable 5.2 and Mythos 5.2: Pricing, Access and API Migration Changes
- Gemini Omni 1.2 Flash: longer video, lower pricing and the limits to check
- Muse Spaces turns a prompt into a personal workspace inside Meta’s AI app
- Inkling 2 Small arrives on Tinker, with downloadable weights still planned
- BioBasicsBench asks whether AI can read the evidence in the lab
- Images 2.6 adds GIFs and new editing controls: what changes for creators
- Sarvam Prism: a decision model that scores choices instead of writing replies
- LongCat-2.5 reaches the API: coding-agent support, limits and introductory pricing
- Gemini adds in-chat file editing: PDFs on mobile, documents and slides on the web
- OpenAI backs a shared AI-training signal: what website owners should know
- Pika H3 Flex: what creators get locally, and what stays in the cloud
- Gemini 4 Carbon: limited access, launch pricing and the evidence behind its security claims
- Rime launches Prism: a voice AI that lets applications check replies before speech
- Genie 4 adds sound to interactive AI worlds: access, limits and what to watch
- GPT-6.1 Terra launches: access, pricing and what to test before automating screen work
- Mercury 3 arrives: launch pricing, API access and the limits of a speed headline
- NVIDIA releases Bolt E4B: what its 4GB target means for local AI assistants
- Kimi K3.1 Pro launches: access, API pricing and when Highspeed is worth testing
- 1Password for Agents launches: approved logins, rollout dates and what access still means
- Claude Duet arrives: who can use it, usage limits and a useful first test
- Mistral launches Voxtral 3: voice agents, pricing and what to check before taking calls
- MiniMax M3.1 Flash launches: pricing, access and the cost of a finished task
- DeepSeek opens its search API: pricing, Flash 1017 and what to test
- Reflection releases Beam: open weights, coding agents and what to test first
- Meta Muse Spark Open explained: specs, pricing and what it means for you
- Grok Voice 3.0 launches: what businesses should test before answering calls
- Gemini 4 Argon opens to paid users: access, pricing and what to test
- Meta launches Muse Wonder: what its memory and focus mean for AI agents
- Microsoft-Decision-1 launches: cheaper AI for routing and agent checks
- Britain’s AI Growth Zones face an infrastructure test
- The EU’s voluntary AI code: what its three chapters cover
- The NIST AI framework remains a useful risk-management lens
- Manchester researchers use Earth-2 for urban air-quality forecasts
- Perplexity’s Windows rollout brings local RTX agents into view
- Singapore’s agentic AI framework puts accountability in focus
- A UK commission sets out a blueprint for healthcare AI
- AI literacy becomes an organisational planning question
- A new alliance looks at flexible AI data-center demand
- Open-source AI research reaches pediatric cardiac care
- IndiaAI’s July update maps a national rollout
- OpenAI proposes a framework for reporting model misalignment
- ChatGPT’s financial-services push turns to professional workflows
- Microsoft’s education commitment puts AI skills on the agenda
- Microsoft 365 G7: AI tools for US government work
- NVIDIA’s new inference results invite a closer benchmark reading
- NVIDIA’s infrastructure update turns attention to tokens per watt
- Nineteen EU countries design a joint AI project
- OpenAI’s data-work announcement puts everyday analysis in focus
- OpenAI’s Agents API puts orchestration in the spotlight
- Industrial AI turns from demonstration to daily operation
- Lyria’s latest update raises the music-workflow question
- AI-assisted contrail research expands to long-haul flying
- Google’s AI Economy Atlas adds a new lens on adoption
- Claude Haiku 5.5 launches: lower AI costs, with a catch for longer prompts
- Inside the storage problem behind a large AI service
- GPT Live 1 brings another real-time API into view
- Microsoft Marketplace adds an intelligent-discovery layer
- Mapping methane from space: AI meets an invisible problem
- Antigravity explores teamwork between coding agents
- Connecting AI usage to business value becomes a reporting task
- Mistral’s new funding round sharpens Europe’s AI ambitions
- Mistral’s agentic search pushes beyond a results page
- Shieldstral adds another layer to model safeguards
- Google Research applies AI to wildfire tracking
- Google updates its real-time audio building blocks
- OpenAI reframes the advertising conversation around AI
- Mistral and Cloudera connect models with enterprise data
- Mistral and HUMAIN outline a new regional partnership
- Gemini arrives on Windows with desktop ambitions
- A model-hardware standard moves into research preview
- Anthropic expands its support for scientific work
- Mistral and Mozilla bring AI into the browser conversation
- Mistral adds European regional inference options
- Voxtral TTS extends Mistral’s speech stack
- Anthropic’s September threat report puts misuse monitoring in focus
- Enterprise Frontier Safeguards brings deployment controls into view
- Alignment and security research gets a joint agenda
- Anthropic separates general and trusted access in its 5.1 release
- Claude Opus 5: reading a frontier-model announcement carefully
- Anthropic explains its text-watermarking rollout
- Google announces Gemini 4 Argon: what the limited rollout means
Tools
- Otter: organize spoken information for follow-up
- Fireflies: review meeting records before acting on them
- Fathom: turn meeting highlights into an action list
- Glean: ask workplace questions against company knowledge
- Cohere: approach enterprise retrieval as an evaluation problem
- Hugging Face: inspect a model before trying it
- Ollama: begin a local-model experiment
- LM Studio: evaluate a model on your own computer
- DeepL: build a translation workflow with human review
- Udio: explore a musical direction without losing the brief
- Notion AI: turn workspace knowledge into useful answers
- Coda AI: connect a document workflow to structured information
- Zapier Agents: define the job before automating it
- Make: map the decisions inside an AI automation
- n8n: build an inspectable AI workflow
- Lindy: scope an assistant to a repeatable business task
- Gumloop: connect AI steps into an understandable process
- Granola: turn a conversation into useful meeting notes
- Pika: test a short visual idea with a clear shot plan
- Luma Dream Machine: develop a scene through visual references
- Sora’s shutdown: preserve your work and rebuild a video workflow
- HeyGen: prepare a clear presenter-led video
- ElevenLabs: design a voice workflow with review checkpoints
- Descript: edit a spoken story through its transcript
- Figma Make: explore an interactive design concept
- Midjourney: build a consistent visual direction
- Recraft: develop a coherent illustration family
- Veo: write a production brief for generated video
- Synthesia: structure a training video before production
- Suno: brief a musical concept with clear intent
- Bolt: build a focused browser-based prototype
- Adobe Firefly: direct a visual generation workflow
- Leonardo: organize image experiments around one art direction
- Cursor: review an AI-assisted code change with context
- Devin Desktop (formerly Windsurf): keep AI coding reviewable
- v0: turn interface requirements into a working starting point
- Lovable: prepare a useful brief for an app builder
- Canva AI: turn one campaign idea into a visual system
- Ideogram: plan typography-led image concepts
- Runway: storyboard before generating a video
- Gemini Notebook (formerly NotebookLM): organise a source-based briefing
- Codex: plan a reviewable software change
- Replit Agent: take a small app from idea to prototype
- ChatGPT: build a research-to-draft workflow
- Microsoft Copilot: start with a defined workplace task
- Perplexity: explore a topic through linked sources
- Claude Code: scope a coding agent to a concrete task
- GitHub Copilot: move from suggestions to reviewed code
- Gemini: connect multimodal work to a clear question
- Claude: turn a long document into a structured review
- Approving AI agent actions with HOL Guard: What to check first
Compare
- ChatGPT Astra vs Gemini Argon vs Claude Fable: Which AI should you choose?
- No-code vs code-based AI workflows: choose the right ownership
- Accuracy tests vs preference tests: measure the right kind of quality
- Voice vs text interfaces: fit the interaction to the setting
- Reference images vs text prompts: control a visual direction
- Consumer vs enterprise AI plans: compare the operating requirements
- Managed cloud vs self-hosting: compare the operational burden
- Text-to-video vs image-to-video: choose the right starting point
- Chat interfaces vs embedded copilots: meet users inside the work
- Seat-based vs usage-based AI: compare the cost structure
- Agents vs fixed workflows: where should decisions happen?
- Single-agent vs multi-agent systems: compare coordination costs
- Human review vs autonomous execution: choose by consequence
- AI avatars vs voiceovers: match format to the audience
- Prompt libraries vs custom assistants: share instructions or behavior
- API access vs a chat subscription: choose the delivery model
- Reasoning vs fast-response models: when the wait is worthwhile
- Vector vs keyword search: compare retrieval signals
- Long context vs retrieval: choose how evidence reaches the model
- Real-time vs batch AI: compare urgency and cost
- Otter vs Fireflies: choose a meeting knowledge workflow
- DeepL vs an AI assistant: evaluate translation in context
- Open weights vs closed models: compare the responsibilities
- RAG vs fine-tuning: knowledge access or behavior change?
- Tool calling vs prompt-only work: when actions need an interface
- Structured output vs free text: choose the right contract
- n8n vs Gumloop: compare control and workflow building
- Ollama vs LM Studio: two routes into local AI
- Hosted AI vs local AI: where should the model run?
- Midjourney vs Ideogram: evaluate image style and lettering
- HeyGen vs Synthesia: evaluate presenter-led video
- Notion AI vs Coda AI: knowledge workspace or structured document
- Granola vs Fathom: compare notes and follow-up
- Glean vs a custom knowledge assistant: buy or build retrieval
- Large vs small models: match capacity to the task
- Canva AI vs Adobe Firefly: campaign production or asset creation
- Runway vs Luma: choose around the shot you need
- Suno vs Udio: compare a music-generation workflow
- Claude vs Gemini: evaluate document-heavy work
- Cursor vs GitHub Copilot: compare coding workflows
- Replit Agent vs Lovable: choose a prototype workflow
- Recraft vs Leonardo: compare illustration workflows
- Descript vs ElevenLabs: transcript editing or voice production
- Zapier vs Make: choose an AI automation approach
- ChatGPT vs Gemini: compare the workflow around the answer
- Gemini Notebook vs Perplexity: curated sources or wider discovery
- Claude Code vs Codex: compare coding agents on bounded work
- ChatGPT vs Claude: choose around your actual tasks
- Perplexity vs ChatGPT: search-led research or a broader assistant
- Devin Desktop vs Cursor: compare the current coding workflow
- v0 vs Bolt: interface generation or a broader build loop
Founders
- Flo Crivello: Building assistants around repeatable business work
- Jonathan Ross: Designing infrastructure around AI inference
- Chris Pedregal: Designing AI around the meeting notebook
- Mikey Shulman: Making music creation more accessible
- Dani Yogatama: Building models for more than one kind of input
- Bhavish Aggarwal: Building an AI company for Indian use cases
- Krish Ramineni: Turning meetings into searchable knowledge
- Sam Liang: Making spoken information easier to use
- Andrew Mason: Rethinking editing through the transcript
- Andrew Feldman: Rethinking the hardware used to train AI
- Yi Tay: Bringing research experience into a multimodal startup
- Amnon Shashua: Connecting AI research and company formation
- Pratyush Kumar: Taking language research toward practical deployment
- Alexandr Wang: Building infrastructure around training data
- Ashish Vaswani: From model architecture to company building
- Noam Shazeer: Designing conversational AI as an experience
- Daniel De Freitas: Building a product around open-ended conversation
- Yoav Shoham: Turning language research into AI products
- Vivek Raghavan: Building for India's languages and institutions
- Lukas Biewald: Making machine-learning work easier to inspect
- Daniela Amodei: Building the organization around an AI research lab
- David Luan: Building toward AI that works with software
- Cristóbal Valenzuela: Bringing machine learning into creative production
- Victor Riparbelli: Rethinking how organizations make video
- Aidan Gomez: Bringing language models into enterprise workflows
- Jerry Liu: Connecting models with external data
- Andrew Ng: Building an ecosystem for applied AI learning
- Lucy Guo: A co-founder perspective on AI infrastructure
- Niki Parmar: Connecting foundational research and enterprise tools
- Demi Guo: Making video generation a creative interface
- Mati Staniszewski: Turning voice research into a product platform
- Harrison Chase: Building the tools around language-model applications
- Arthur Mensch: Building a European AI model company
- Varun Mohan: The founder journey behind an AI coding platform
- Eric Simons: Bringing the development environment into the browser
- Amit Jain: Working toward a richer understanding of the visual world
- Joshua Xu: Building tools for video communication
- Clément Delangue: Building a shared home for machine learning
- Michael Truell: Building AI into the code editor
- Anton Osika: Turning a natural-language brief into a product
- David Holz: Building a community around visual imagination
- Sam Altman: Building a platform around general-purpose AI
- Fei-Fei Li: Building toward spatial intelligence
- Aravind Srinivas: Reimagining the search-and-answer experience
- Amjad Masad: Lowering the barrier from idea to software
- Guillermo Rauch: Connecting developer infrastructure and AI interfaces
- Demis Hassabis: Connecting scientific ambition and machine learning
- Ilya Sutskever: A company organized around one technical ambition
- Dario Amodei: Turning research priorities into an AI company
- Mira Murati: Designing a new AI research organization
Learn
- Responsible adoption: start with one measurable pilot
- AI for research: build a traceable reading workflow
- AI for design: use constraints to improve exploration
- AI for meetings: move from notes to accountable actions
- Code-switching: testing AI across mixed-language conversations
- AI for spreadsheets: check formulas and assumptions
- AI for sales: improve preparation without inventing personalization
- AI literacy: ask better questions about an AI product
- Agent memory: separate stored facts from conversation history
- Prompt injection: treating outside content as untrusted
- Accessibility: design AI output that more people can use
- AI for writing: separate factual review from style
- AI for customer support: make escalation part of the product
- AI for education: support practice and feedback
- Batch processing: handling AI work that can wait
- Agent handoffs: design a useful transfer of responsibility
- AI bias: look at who is missing from your evaluation
- Small language models: when a smaller model is enough
- Quantization: shrinking models without ignoring quality
- AI cost: count the whole workflow, not one response
- Multi-agent systems: when collaboration adds value
- AI privacy: map the data before choosing a workflow
- Multilingual AI: evaluate meaning, not just translation
- Voice synthesis: designing a clear and responsible voice
- Local AI: what running a model on your computer means
- Model latency: measure the wait that users actually feel
- Evaluations: build a test set before choosing a model
- Multimodal AI: working across text, images and sound
- AI video: plan a shot before generating a clip
- Open weights: access, licensing and practical trade-offs
- Distillation: teaching a smaller model from a larger one
- Caching: reuse work when the underlying inputs stay valid
- Model hallucinations: why plausible answers can be wrong
- Human oversight: choosing the right approval points
- Diffusion models: a plain-language guide to denoising
- Embeddings: turning meaning into searchable numbers
- Fine-tuning: when model customization makes sense
- Tool calling: how models connect to useful actions
- Grounding: connecting an answer to its evidence
- Image generation: guide composition before adding detail
- Speech recognition: turning audio into workable text
- Context windows: why more space is not perfect memory
- RAG: giving an assistant the right source material
- System instructions: setting the boundaries of an assistant
- Large language models: what they actually learn
- Tokens: the units behind every AI conversation
- Vector search: finding ideas beyond exact words
- Prompting: write a useful brief before asking for output
- Structured outputs: making AI answers usable by software
- AI agents: from a goal to a completed task