The Most Common Mistake People Make When Building AI Agents
Key Takeaways
Anthropic researcher Erik Schluntz discusses the common mistake of creating detailed prompts but neglecting to provide proper documentation and tooling for AI agents, highlighting the importance of prompt engineering in tool descriptions.
Full Transcript
people will put a lot of effort into creating these really beautiful detailed prompts and then the tools that they make to give the model are sort of these incredibly bare bones uh like you know no documentation fun like the parameters were named A and B and it's kind of like oh like an an engineer wouldn't be able to like work with this as if this was a function they had to use because there's no documentation like how can you expect Paul to use this as well so kind of it's like that lack of like putting yourself in the model shoes I think a lot of people when they start trying to use tool use and function calling. um they kind of forget that they have to prompt as well and they they think about the model just as this you know more classical programming system but it is still a model and you need to be prompt engineering in the descriptions of your tools themselves.
Original Description
Anthropic researcher Erik Schluntz shares the most common mistake he's observed of people building AI agents.
Watch on YouTube ↗
(saves to browser)
Sign in to unlock AI tutor explanation · ⚡30
Playlist
Playlist UUrDwWp7EBBv4NwvScIpBDOA · Anthropic · 50 of 60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
▶
51
52
53
54
55
56
57
58
59
60
Quick tips for Claude: Long context file uploads
Anthropic
Inside our first Anthropic Hackathon, San Francisco
Anthropic
Long inputs, multi-step output with Claude
Anthropic
Coding with Claude
Anthropic
Behind the prompt: Prompting tips for Claude.ai
Anthropic
Robin AI, powered by Claude
Anthropic
Claude 3 Opus as an economic analyst
Anthropic
Claude 3 Sonnet as a language learning partner
Anthropic
Claude 3 Haiku turns thousands of physical documents into structured data
Anthropic
Claude 3 Haiku for instant customer service
Anthropic
Claude 3 Haiku for fast document analysis
Anthropic
Tool use with the Claude 3 model family
Anthropic
Coming soon to the Team plan on Claude.ai
Anthropic
Introducing the Claude iOS app
Anthropic
Claude is now available in Europe
Anthropic
What is interpretability?
Anthropic
What should an AI's personality be?
Anthropic
Scaling interpretability
Anthropic
Claude 3.5 Sonnet for sparking creativity
Anthropic
Claude 3.5 Sonnet for vision
Anthropic
Claude 3.5 Sonnet as a writing partner
Anthropic
Claude 3.5 Sonnet for agentic coding
Anthropic
Shareable Projects in Claude
Anthropic
Evaluate prompts in the Anthropic Console
Anthropic
Shareable Artifacts in Claude
Anthropic
How we built Artifacts with Claude
Anthropic
Wedia advances digital asset management with Claude
Anthropic
AI prompt engineering: A deep dive
Anthropic
AI Prompt Engineering 101: Explained
Anthropic
Ancient Wisdom, Modern AI?
Anthropic
AI's Greatest Challenge: You?
Anthropic
AI Prompts That Drive Growth
Anthropic
Tips For Better Results With AI
Anthropic
AI, policy, and the weird sci-fi future with Anthropic’s Jack Clark
Anthropic
European Parliament expands access to their archives with Claude in Amazon Bedrock
Anthropic
Claude | Computer use for automating operations
Anthropic
Claude | Computer use for orchestrating tasks
Anthropic
Claude | Computer use for coding
Anthropic
Asana supercharges work management with Claude
Anthropic
What do people use AI models for?
Anthropic
Alignment faking in large language models
Anthropic
Building Anthropic | A conversation with our co-founders
Anthropic
How difficult is AI alignment? | Anthropic Research Salon
Anthropic
Tips for building AI agents
Anthropic
Claude 3.7 Sonnet with extended thinking
Anthropic
Introducing Claude Code
Anthropic
Advice For Building AI Agents
Anthropic
The Two Most Useful Applications of AI Agents
Anthropic
Defending against AI jailbreaks
Anthropic
The Most Common Mistake People Make When Building AI Agents
Anthropic
Controlling powerful AI
Anthropic
How Intercom is redefining customer support with Claude
Anthropic
Tracing the thoughts of a large language model
Anthropic
Introducing Claude for Education
Anthropic
Could AI models be conscious?
Anthropic
Lessons on AI agents from Claude Plays Pokemon
Anthropic
The Societal Impacts of AI
Anthropic
What Does AI Mean for the Future of Work?
Anthropic
Understanding AI Agents...Through Pokémon
Anthropic
What Pokémon Teaches Us About Building With AI
Anthropic
More on: Prompt Craft
View skill →Related AI Lessons
⚡
⚡
⚡
⚡
My agent kept reading data it wasn't allowed to. The prompt was never going to stop it.
Dev.to AI
8 Must-Know AI Chatbot Tools That Actually Help Small Businesses
Dev.to AI
Agent-Ready Commerce, Part 9: Evidence and Audit Are Part of the Product
Dev.to AI
Agent-Ready Commerce, Part 8: Generated Claims Need Review, Evidence, and Expiry
Dev.to AI
🎓
Tutor Explanation
DeepCamp AI