I use Anthropic's Claude AI tools for very different jobs: How to pick between models, Code, and Cowork

Summary

Anthropic's Claude AI offers distinct tools: Claude.ai (chatbot), Claude Code (coding agent), and Claude Cowork (productivity agent). LLMs like Haiku, Sonnet, Opus, Fable, and Mythos represent different engine sizes, with Fable and Mythos being the most powerful. Agents can perform complex tasks, plan, and execute sequences, significantly speeding up development for experienced users.

Key Takeaways

1. Claude's lineup spans chatbots, agents, and AI models.
2. Code builds software. Cowork handles broader computer tasks.
3. Agents can be force multipliers but require close supervision.

Statement Breakdown

• Claimed Facts: 60% of statements the article presents as facts
• Opinions: 30% of statements classified as editorial or subjective
• Claims: 10% of statements surfaced for additional reader evaluation

Credibility & Bias Reasoning

Credibility assessment: The article provides a detailed comparison of AI tools, using analogies to explain complex concepts. It cites specific product names and features, and the author's background as a former computer science professor and engineering manager lends authority. However, some claims about market valuations and AI capabilities are presented without direct external verification.

Bias assessment: Enthusiastic Tech Advocate. The author expresses significant enthusiasm for AI technologies, particularly Anthropic's Claude tools. While aiming for objectivity, the language often highlights the transformative potential and impressive capabilities of these tools, framing them as powerful solutions for various tasks.

Note: This article offers a detailed look at AI tools with personal anecdotes. While informative, some claims about market valuations and AI capabilities may benefit from cross-referencing with other sources.

Credibility flag: Enthusiastic, but verify

Claimed Facts (7)

• This presents a specific financial claim about Anthropic's market standing.
• This states a specific event regarding the US government's ban on a particular AI model.
• This statement categorizes specific Anthropic models as LLMs and draws an analogy to engine sizes.
• This provides a fundamental distinction between chatbots and agents, serving as a foundational explanation.
• This statement categorizes specific AI tools as chatbots.
• This asserts Claude Code's prominence among coding agents and notes the trend among other providers.
• This states that both Anthropic and OpenAI offer productivity-focused AI agents.

Opinions (9)

• This is a rhetorical question that expresses a subjective view on the public perception of Anthropic's products.
• The description 'very nice' is a subjective assessment of the car.
• The term 'monster' is a subjective and evocative description of the engine's power.
• These are personal preferences regarding the car's performance and sound.
• The assessment 'nice' and 'not too powerful' are subjective judgments.
• The evaluation 'Good, but not great' is a subjective opinion on the performance of these models.
• This statement reflects the author's personal experience and perceived necessity for guidance, which is subjective.
• This is a simile used to describe the author's experience, which is a subjective comparison.
• The terms 'powerful' and 'fundamentally freeing' express the author's personal feelings and assessment of the experience.

Claims (6)

• While market valuations can be reported, the term 'astonishing' and the specific figure without direct source attribution for the valuation itself makes it a claim requiring verification.
• The phrase 'just too smart and risky for our own good' is an interpretation of the reason for the ban, potentially an oversimplification or emotional framing.
• This is a sensational claim about revenue generated by Claude Code, presented without supporting financial data or a clear explanation of how the author's app demonstrates this.
• The phrase 'vibe code' is informal and lacks a clear technical definition, and the statement implies a subjective and potentially exaggerated experience of the effort involved.
• Similar to the previous point, 'vibe code' is unclear, and the dramatic time saving is presented without detailed methodology, making it a potentially exaggerated claim.
• The phrase 'earned a spot in my too' is incomplete and lacks context, making the claim about Claude Cowork's performance vague and unsubstantiated.

Key Sources

• ZDNET — Media Outlet
• David Gewirtz — Author, Former Computer Science Professor, Product Marketing Director, Engineering Manager
• Anthropic — AI Company
• OpenAI — AI Company
• Google — Technology Company