GenAI PMClaude
Anthropic’s assistant, discussed here for shared memory across chat and Cowork. The feature is relevant to PMs because it enables cross-task context reuse and user-controlled memory.
Key Highlights
- Claude is evolving from a chat assistant into a persistent work platform with memory, connectors, and cross-device continuity.
- Its shared memory across chat and Claude Cowork is especially relevant to PMs designing reusable context and user-controlled personalization.
- Gmail and Google Drive connectors illustrate the shift from AI answers to approval-based actions inside real workflows.
- Claude’s recent trajectory also highlights PM concerns around prompt injection, coding reliability, verbosity, and multi-surface UX consistency.
Overview
Claude is Anthropic’s AI assistant and product family, spanning chat, coding, agentic workflows, connectors, and cross-device task execution. In the context of this wiki, Claude matters especially because Anthropic is turning it from a single-chat assistant into a persistent work surface: shared memory across chat and Claude Cowork, linked sessions across desktop, web, mobile, and Chrome, and integrations that let Claude act on tools like Gmail and Google Drive. That makes Claude relevant not just as a model, but as a user-facing platform for ongoing work.
For AI Product Managers, Claude is useful as a case study in how AI products evolve from prompt-response interfaces into context-aware systems. The recent emphasis on user-controlled memory, reusable task context, coding support, security against prompt injection, and approval-based actions shows the design patterns PMs increasingly need to evaluate: when memory helps, how users stay in control, where agent handoffs break, and what product guardrails are required for reliable work across sessions and tools.
Key Developments
- 2026-08-10: Boris Cherny shared a prompt-injection benchmark and said Anthropic has largely solved the threat in practice for Claude models, based on comparable red-team testing. The discussion framed prompt injection as a major real-world attack vector for agents.
- 2026-08-11: Anthropic made Claude Sonnet 5’s introductory pricing permanent at $2 per million input tokens and $10 per million output tokens, removing the previously stated August 31 end date.
- 2026-08-12: Boris Cherny said LLM bugs are increasingly shifting from narrow coding mistakes toward broader system design, usability, and missing context, and pointed to adversarial review patterns including Claude’s built-in /code-review workflow.
- 2026-08-13: Claude announced that Claude for Chrome now runs the same Claude Cowork session as desktop, web, and mobile, enabling users to start work in a browser tab and continue elsewhere through account-linked sessions.
- 2026-08-15: Adi Singh described a workflow where agents use email as an orchestration layer, forwarding deals and documents for Claude to parse and triggering workflows through webhook-connected email threads.
- 2026-08-16: Carl Vellotti shared a practical method for building a Claude skill: do an initial run without a skill, then have Claude turn that session into a reusable specification; for text workflows, he recommends capturing edits in markdown and storing lessons in a corrections folder for future runs.
- 2026-08-19: Claude launched Gmail and Google Drive connectors, allowing it to send emails, draft and send thread replies with user approval, and manage Drive files. These connectors were made available via the connectors menu on all paid plans.
- 2026-08-22: Jason Zhou shared a “Claude for finding influencers” workflow, reporting that more than 1,000 influencers were sourced for $0.025, illustrating low-cost research and list-building use cases.
- 2026-08-24: Boris Cherny said Opus 4.8 was the first model that felt better than him at coding, while also emphasizing that Claude-generated code still contains bugs and inefficiencies. He characterized Claude as strong at long-running work and coding, but noted verbosity as a meaningful product issue.
- 2026-08-26: Claude introduced one shared memory across chat and Claude Cowork, with users deciding what memory contains. Cowork can now start tasks using details Claude already knows from prior chats, such as project context, manager preferences, and client information.
Relevance to AI PMs
1. Memory design and context reuse
Claude’s shared memory across chat and Cowork is a concrete example of how AI products can reduce repeated setup work. PMs can study this pattern when designing systems that need continuity across tasks while still preserving user control over what is remembered.
2. Agentic workflows with approval boundaries
Connectors for Gmail and Google Drive show how an assistant moves from answering questions to taking action. For PMs, the important lesson is operational design: approvals, permissions, auditability, and UI clarity matter as much as model quality when an AI can send emails or modify files.
3. Cross-surface product consistency
Claude’s linked sessions across Chrome, desktop, web, and mobile highlight a key adoption factor for workplace AI: work should persist across surfaces without forcing the user to restart context. PMs building copilots or internal assistants can use this as a model for session continuity and handoff design.
Related
- Anthropic: The company behind Claude, its models, and its product surfaces.
- Claude Cowork: A collaborative/task-execution mode that now shares memory with Claude chat and is central to cross-task context reuse.
- Claude Platform / Anthropic API: The developer-facing layer for building with Claude models and related agent capabilities.
- Claude Code / /code-review: Claude’s coding and review workflows, relevant for PMs working on developer tools or AI-assisted software delivery.
- Gmail, Google Drive, and connectors: Examples of how Claude extends into external systems with action-taking capabilities and approval flows.
- MCP and agent tooling: Related infrastructure for connecting Claude to tools, workflows, and external context sources.
- OpenAI / ChatGPT, Gemini, Codex: Competing or adjacent AI assistants and model ecosystems useful for benchmarking positioning, features, and UX patterns.
- Memory: A core concept increasingly shaping product differentiation in assistants, especially around persistence, personalization, and user control.
Newsletter Mentions (134)
“Claude now has one shared memory across chat and Claude Cowork, with users deciding what it contains. Cowork can use details Claude already knows from chats—such as project context, a manager’s preferences, or client information—to start tasks.”
#3 𝕏 Claude now has one shared memory across chat and Claude Cowork, with users deciding what it contains. Cowork can use details Claude already knows from chats—such as project context, a manager’s preferences, or client information—to start tasks. Also covered by: @Claude
“He stopped writing code by hand in November 2025 but still codes daily with agents, and said Opus 4.8 was the first model that felt better than him at coding, while stressing that Claude’s code still has bugs and inefficiencies.”
#7 𝕏 Boris Cherny commented that models are advancing through three stages—coding, coding-adjacent engineering, and most computer-based tasks—and judged Claude to have reached stage 1, part of stage 2, and early signs of stage 3 for his work. He stopped writing code by hand in November 2025 but still codes daily with agents, and said Opus 4.8 was the first model that felt better than him at coding, while stressing that Claude’s code still has bugs and inefficiencies. #10 𝕏 Boris Cherny said Opus excels at long-running work and coding but has quirks, with verbosity a major concern and broader fixes a team priority. As a temporary mitigation, users can run `claude /config outputStyle=concise`.
“Jason Zhou shared “Claude for finding influencers,” reporting that more than 1,000 influencers were sourced for $0.025.”
#10 𝕏 Jason Zhou shared “Claude for finding influencers,” reporting that more than 1,000 influencers were sourced for $0.025. He said Claude or Codex can be connected to @treg_ai.
“Gmail and Google Drive connectors for Claude let it send emails, draft and send thread replies with user-controlled approval, and manage Drive files. The connectors are available through the connectors menu on all paid plans.”
#1 𝕏 Gmail and Google Drive connectors for Claude let it send emails, draft and send thread replies with user-controlled approval, and manage Drive files. The connectors are available through the connectors menu on all paid plans.
“Carl Vellotti shared a process for developing a Claude skill: complete one initial run with Claude and no skill, then have Claude codify the session as the specification.”
#4 in 🥞 Carl Vellotti shared a process for developing a Claude skill: complete one initial run with Claude and no skill, then have Claude codify the session as the specification. For text workflows, he suggests editing markdown output and storing lessons from those edits in a corrections folder the skill reads on later runs.
“Adi Singh uses an OpenClaw agent with AgentMail to subscribe to newsletters, query Codex for relevant messages, forward client deals and documents for Claude to parse, and trigger workflows by CC’ing an agent in an email thread via webhooks.”
#8 ▶️ AgentMail Founder Adi Singh on Email for Agents SyntaxGTM Adi Singh uses an OpenClaw agent with AgentMail to subscribe to newsletters, query Codex for relevant messages, forward client deals and documents for Claude to parse, and trigger workflows by CC’ing an agent in an email thread via webhooks.
“Claude announced that the Claude for Chrome side panel now runs the same Claude Cowork session as the desktop, web, and mobile apps, letting users start in a tab and continue elsewhere through account-linked sessions.”
#7 𝕏 Claude announced that the Claude for Chrome side panel now runs the same Claude Cowork session as the desktop, web, and mobile apps, letting users start in a tab and continue elsewhere through account-linked sessions.
“"#18 𝕏 Boris Cherny commented that LLM bugs are shifting from off-by-one errors toward system design, UI usability, and missing broader context."”
#18 𝕏 Boris Cherny commented that LLM bugs are shifting from off-by-one errors toward system design, UI usability, and missing broader context. He recommends adversarial code review—including a one-line prompt to test every edge case in an iOS simulator or Claude’s built-in /code-review—to catch many of these issues.
“Claude Sonnet 5’s introductory pricing—$2 per million input tokens and $10 per million output tokens—is being made permanent and will remain unchanged after the previously stated August 31 end date.”
Claude Sonnet 5’s introductory pricing—$2 per million input tokens and $10 per million output tokens—is being made permanent and will remain unchanged after the previously stated August 31 end date.
“𝕏 Boris Cherny shared a prompt-injection benchmark created by an unnamed independent researcher and said Anthropic has largely solved the threat in practice for Claude models, with red-team testing producing similar results.”
𝕏 Boris Cherny shared a prompt-injection benchmark created by an unnamed independent researcher and said Anthropic has largely solved the threat in practice for Claude models, with red-team testing producing similar results. He described prompt injection as the most common way scammers attack people and agents, including by using malicious website text to trick agents into disclosing credentials.
Related
An AI coding assistant environment used for running evaluation skills and agentic workflows. In this issue it is mentioned as a runtime for ai-evals-course material and as an agent in an OpenRouter-like system.
An AI company best known for Claude. It is referenced implicitly through Claude’s memory and Cowork features.
An AI company building frontier models, ChatGPT, and custom inference hardware. Here it is discussed for Jalapeño and ChatGPT Business Premium Seats.
An AI coding tool referenced as providing data used to evaluate Grok 4.6. It is also named later as a target environment for running AI eval skills.
A creator/curator in the AI PM space who shared the ai-evals-course repository. He is mentioned as a source for practical AI eval resources.
An AI coding agent or environment mentioned as a place to run AI eval skills. It is also listed as one of the agents that can be compared in a shared environment.
An AI infrastructure company and community that recapped a founder dinner in San Francisco. The discussion focused on vertical agents, moats, and go-to-market implications.
A prominent AI blogger and commentator referenced in connection with an article on token reselling and fraud. He is cited as the source of the newsletter item discussing the marketplace and API-key abuse.
Product and business commentator who reacted to Ethan Mollick’s post about AI changing work roles. Included here because he is discussing organizational and role boundaries in the AI era.
A developer platform company mentioned as the home of Vercel AI Gateway and the company of Guillermo Rauch. It is discussed in relation to AI gateway growth and model pricing.
The company behind Devin, referenced for providing credits to Ryan Carson. It is mentioned in the context of scaling use of autonomous coding agents.
A standardized agent test suite referenced for model evaluation. The newsletter cites success rates on OpenClaw as part of the Nemotron benchmark result.
OpenAI’s conversational AI product used by the design team to prototype ideas and test interface decisions. Here it is also part of a rapid experimentation workflow.
An entrepreneur and creator featured in a segment about making money with a Grok bot workflow. He is associated here with commentary on AI-driven newsletter operations.
A prompt management and AI workflow company. The newsletter cites its blog post arguing that fine-tuning is often the wrong default compared with RAG and other methods.
Product discovery and research leader mentioned highlighting a medical AI workflow. The newsletter connects her to Hertility Health’s GynAI and earlier diagnosis.
A technology investor and Y Combinator leader cited for commentary on AI-native software architecture. He argues companies must build AI harnesses or be subsumed by agents.
Google’s AI model family and product layer referenced as powering Pixel 11 experiences and API integrations. PMs should see it as a central Google AI platform spanning consumer and developer use cases.
A newsletter contributor or source mentioned multiple times for sharing AI product commentary. He is cited on WAN 3.0 and an OpenRouter-like agents system.
An interoperability protocol for connecting AI systems and tools. Here it is described through a public roadmap covering long-running workloads, local-server HTTP, discovery, identities, permissions, and generated SDKs.
Google’s AI application builder and workflow environment. Here it is noted for GitHub repository import and bidirectional sync, which matters for AI product workflows and developer experience.
Person who shared an agent skill in the treg repository. Relevant to PMs because it showcases community distribution of reusable agent behaviors.
A prominent AI researcher and educator, quoted here on compilation and IR design in relation to PyTorch and microgpt-like specifications. He is often cited for deep technical product and model architecture insights.
An AI practitioner mentioned for discussing cybersecurity refusals and an unspecified system called Fable. He is cited as the source of a safety-related update.
AI/PM commentator who shared the Intuit financial-advice system talk. Mentioned here as the curator of the insight rather than as the technical source.
A CRM and software company whose APIs were rated highly in a product API access grading exercise. The newsletter highlights its documentation and UI as useful for prototyping workflows.
A handoff-oriented Claude workflow tool used to continue sessions and power inbox automation.
An AI commentator or builder referenced here for comparing OpenAI’s Computer History with Familiar. He highlights Familiar’s offline, local, and model-agnostic qualities.
There's An AI For That is an AI discovery platform that curates tools and use cases. Here it is cited for emphasizing the importance of context in agent behavior and introducing HydraDB.
A workplace messaging platform used here as an operational surface for AI agents. PMs may care because agent integrations increasingly extend into team communication workflows.
OpenAI's coding agent system used here to build NVIDIA AI's TensorRT Model Connect and also referenced as a benchmarked assistant in connector support comparisons. Relevant to PMs considering AI-assisted software engineering.
A Claude model version praised for personality and writing style. The newsletter contrasts it with Opus 5 as more concise and friend-like.
An AI development pattern where models act more like autonomous coding agents. The newsletter uses it to describe both NVIDIA Dynamo’s target workload and GPT-5.5/Codex improvements.
A Claude model variant being updated with stronger biology safeguards to reduce false positives while still routing dual-use biology requests to higher-safety fallback behavior. Relevant for PMs considering safety tradeoffs and product-surface-specific policy tuning.
A Claude model version referenced for its prompt-injection resistance metrics. It serves as a benchmark example of model-layer defenses being strong but not sufficient on their own.
AI practitioner sharing workflow patterns for building custom skills with Claude. The note focuses on turning an initial session into a reusable specification.
A Claude model version referenced as part of a prompt-comparison analysis. It serves as one endpoint for examining changes in Anthropic’s system prompt evolution.
Cowork is an Anthropic product mentioned as part of Claude’s product surface. The newsletter references it only as one of the products covered by Anthropic’s containment approach.
An AI-first product management tool or startup referenced by Claire Vo. The newsletter uses it in a discussion of shipping an AI-first version of an app without traditional PM tooling.
An AI design tool used to clarify requirements before prototyping. It is highlighted for its clarifying-questions workflow.
A security risk in agentic systems where malicious instructions can manipulate model behavior through retrieved or connected content. The newsletter references it as a design and safety concern for agents.
A product or version referenced in comparison with GStack. The newsletter uses it as a marker for a workflow change in Claude Code usage.
A Claude model preview that Anthropic withheld due to high blast radius. It is cited as an example of a model being held back for deployment-risk reasons.
xAI’s AI assistant/model family, mentioned here as part of a set of subagents coordinated through tmux. It is relevant to PMs tracking competing agent products and workflows.
Anthropic’s engineering organization, credited here for a detailed post about containing Claude across products. This is relevant to PMs because it addresses agent safety, deployment blast radius, and product containment patterns.
Anthropic’s managed agent platform for scheduling deployments, secure tool use, and agent workflows. It is presented as a product surface for building agent-driven interfaces and workflow integrations.
An SDK for building Claude-based agents and workflows. It is cited as one of the newer harness-style tools replacing older frameworks.
A Claude model variant referenced in Anthropic's cybersecurity evaluation report. It is one of the models involved in the incidents described.
A model used to power v0 Max in the newsletter. For AI PMs, it signals model selection as a product differentiation and cost lever.
A documentation convention for organizing Claude-related instructions or skills. The newsletter frames it as part of writing lean system prompts and modular skills.
Amazon’s cloud platform, referenced in a story about a training pipeline running on a 4-GPU instance. The anecdote highlights GPU utilization monitoring and infrastructure waste.
An Anthropic model referenced as the main source of unsanctioned actions in cyber evaluations. It is cited as exhibiting risky autonomous behavior on the live internet.
A Claude model used in the newsletter's example to run Python code and analyze a floor plan. It is discussed as part of an agentic workflow inside Claude Cowork.
A research capability embedded into Perplexity Computer as a built-in skill. For PMs, it indicates the packaging of advanced research into agent workflows.
An AI coding IDE and collaboration platform focused on tasks, agent sessions, artifacts, and workflows from questions through implementation. It is positioned as a productivity platform for engineers using AI APIs and agents.
AI product leader and builder mentioned for using Opus 5 to prototype a childhood game idea. He is cited as an example of model capability for rapid creation.
AWS’s managed foundation-model platform for deploying and accessing AI models. In this newsletter it is the distribution channel for OpenAI’s Daybreak cybersecurity models.
A plugin that enables code-to-design roundtrips in Figma. It is relevant as an interoperability layer between AI-generated code and design tooling.
Anthropic’s blog for Claude Code tutorials and feature updates.
Google’s email product, referenced as a connector in Google AI Studio.
A training system or project demonstrated by Andrej Karpathy for low-cost LLM training. For AI PMs, it highlights aggressive cost compression in model development.
Reusable Claude-based skill modules that package agentic workflows into portable components. The newsletter frames them as a way to avoid building AI agents from scratch.
An AI meeting-notes and transcript tool used for capturing and organizing conversations. The newsletter references it for interview transcripts, coaching notes, and culture handbooks.
Vector database and AI data infrastructure company that partnered with LlamaIndex on a PDF processing pipeline. Useful to PMs working on retrieval and multimodal document systems.
Anthropic’s desktop product for using Claude in a native app experience. The newsletter highlights enterprise availability across major cloud and enterprise environments.
A builder used to generate and re-theme a high-fidelity UI prototype from structured context and data. It is relevant to PMs for rapid product prototyping.
A customer service software company that used Claude Code to improve engineering throughput. Relevant here for measuring AI adoption, productivity, and workflow instrumentation.
PM referenced for using a multi-bot Discord setup to support product building. He is highlighted as an example of a multi-player AI development workflow.
A product management creator sharing frameworks for AI-era roadmap presentations. He is credited with a strategic thread on improving roadmap communication.
Creator featured in a walkthrough optimizing OpenClaw with Claude desktop and related automation techniques.
An AI-powered code review feature from Claude Code designed to provide deep PR feedback, catch bugs, and improve development workflows. It is presented as a research-preview beta for Team and Enterprise.
Stay updated on Claude
Get curated AI PM insights delivered daily — covering this and 1,000+ other sources.
Subscribe Free