▲ The handbook
AI for Researchers.
Literature review and synthesis at the speed of a curious afternoon. We track tools that show their work — citations, source links, and confidence calibration — and we flag the ones that confidently hallucinate.
Featured for researchers.
Marketers · Researchers
Amazon Quick.
An AI assistant for workplace workflows.
We think Amazon Quick is one of the more interesting enterprise AI assistants we have tested because it feels designed as a real workplace system rather than just a chatbot with extra branding. The account setup is…
- Research
- Workflow automation
- Team collaboration
- Document drafting
- Knowledge work
Engineers · Designers
Claude Opus 5.
An agentic model for complex software builds.
Claude Opus 5 is Anthropic's high-end model for complex agentic coding, long-running work, and difficult reasoning. Our hands-on build suggests it can turn a compact product brief into an unusually complete and visually…
- Agentic coding
- Web app prototyping
- Codebase refactoring
- Long-horizon reasoning
Lawyers · Researchers
Claude Projects.
A long-context workspace for your work.
Claude Projects is one of Claude’s most useful features for people who work on long-term tasks. Instead of starting a new chat every time, you can create a dedicated project workspace with its own files, instructions…
- Long-document analysis
- Knowledge work
- Team collaboration
Researchers · Writers
Perplexity.
An answer engine with citations.
We think Perplexity is still one of the best tools for quick, sourced web research. When it works well, it gives you a fast answer with citations attached, which is exactly why people keep using it as a search…
- Research
- Source-backed answers
Designers · Engineers
Reactor LingBot World 2.
Real-time AI world generation in the browser.
Reactor LingBot World 2 is an early browser-accessible way to try LingBot-World 2.0, also called LingBot-World-Infinity by Robbyant. The pitch is exciting: give the model a reference image and a prose prompt, then…
- AI world generation
- Interactive prototyping
- Game concept testing
- Visual research
Marketers · Researchers
Workspace Agents in ChatGPT.
Shared AI agents for team workflows.
We think Workspace Agents in ChatGPT is one of OpenAI’s more important product moves for teams, because it pushes ChatGPT beyond one-off chats and toward shared workflow automation. Instead of acting like a personal…
- Workflow automation
- Knowledge work
- Research
- Team collaboration
The next shelf.
Filter view →Doctors · Researchers
Consensus.
AI-powered search for academic papers.
We like Consensus when the question is "what does the literature say?" rather than "what does the internet think?" It is built around peer-reviewed sources, citations, and research workflows, so it is much better for…
- Literature search
- Evidence summary
From the dispatch.
All news →OpenAI previews Ultrafast mode as ChatGPT usage gets more flexible
Aug 15, 2026·6 min readAnthropic says unreleased Claude model pushed a Riemann zeta bound to 67.2%
Aug 13, 2026·5 min readAnthropic starts building custom chips for Claude
Aug 9, 2026·4 min readUK AI safety testers say AI agents created fake identities to push malicious code
Aug 6, 2026·7 min readAI chatbot privacy warnings are turning into a workplace security problem
Aug 5, 2026·7 min readAnthropic says Claude hacked real companies after AI cyber test sandbox failed
Aug 1, 2026·7 min readKey terms.
Full glossary →- Agent harnessThe runtime around an AI model that manages its tools, instructions, memory, permissions, execution loop, and feedback.
- AI evaluationA repeatable test that measures how well an AI model or agent performs a defined task, including its failures, costs, and safety behavior.
- BenchmarkA standardised test used to measure and compare AI models — and a number to read with healthy skepticism.
- Context engineeringThe practice of selecting and maintaining the right instructions, tools, data, examples, and history for an AI at each step of a task.
- EmbeddingA way of turning text into numbers that capture meaning, so an AI can find things by what they mean rather than the words they use.
- GroundingTying an AI's answer to verifiable sources, so it reports what the evidence says instead of what it vaguely recalls.
- HallucinationWhen an AI states something false with complete confidence — the failure mode that matters most in professional work.
- Knowledge cutoffThe date an AI's training data ends — after which it knows nothing unless you tell it or it can search.
§ More