Handle AI rate limits with a queue and bounded retries
Read the actual error and limit dimension, pace concurrent work and avoid retry storms that make a temporary limit worse.
OpenAI — Rate limits
LA24X / AI
Practical guides to using AI tools Page 3
Understand AI tools, check their output and build practical workflows with original sources and clear examples.
Search this section ↗Topics & guides
96 topics in this section.
Estimate AI API cost from usage, not the length of the visible answer
Separate input, cached input, output and tool charges, use current rates and include retries in a realistic workload estimate.
OpenAI — API pricing
Route AI tasks by tested requirements instead of model prestige
Define easy and hard task classes, evaluate candidates on the same examples and use a controlled escalation path.
OpenAI — Model selection
A large context window does not guarantee a complete answer
Understand input and output budgets, keep the needed evidence findable and test whether key conditions survive a long conversation.
docs.anthropic.com
Ask AI about a PDF with page-based evidence, not a vague upload
Check whether the file is text or a scan, ask a defined question and verify tables, conditions and references against the original pages.
OpenAI — File inputs
Use AI vision for screenshots and charts, then verify the detail
Provide a readable relevant image, separate what is visible from interpretation and check small text, units and labels yourself.
OpenAI — Images and vision
Transcribe a recording and verify names, numbers and decisions
Prepare usable audio, choose the required transcript format and correct consequential passages against the recording before sharing.
OpenAI — Speech to text
Create useful AI narration by editing the script before the voice
Prepare a speakable script, preview names and numbers and disclose generated narration rather than implying a real recording.
OpenAI — Text to speech
Design a realtime AI voice session around interruptions and privacy
Choose the supported connection, protect session credentials and test stopping, interrupting and recovering from a dropped call.
OpenAI — Realtime API
Combine text, image and audio for one verifiable task
Choose only the media needed, preserve timestamps and source references and check the answer against each input rather than an invented story.
Google — Gemini video understanding
Use Gemini function calls with a validated execution loop
Define a narrow function, validate the requested arguments and user access and return the real tool result in the supported API format.
ai.google.dev
Use Gemini structured output and keep missing data explicit
Select a supported output schema, represent unknown values deliberately and validate the completed response before importing it.
ai.google.dev
Use Gemini context caching only when reuse covers its cost
Identify repeated input, choose implicit or explicit caching where supported and account for cache creation, storage and expiry.
ai.google.dev
Give Claude a long task with a checkable deliverable
Separate evidence, decisions and constraints, specify the output and ask for missing information rather than an invented completion.
docs.anthropic.com
Design Claude tools with narrow inputs and truthful error results
Define tool schemas and access boundaries, return the matching result and test refusal, timeout and continuation paths.
docs.anthropic.com
Check a Hugging Face Space before uploading private data
Identify the app owner and processing route, inspect its privacy explanation and try harmless data before treating a demo as trusted.
Hugging Face — Spaces overview
Run a local AI model and verify that the work stays local
Check model size and runtime support, distinguish local from cloud-backed modes and test a permitted task with resource monitoring.
Ollama — FAQ and configuration
Quantization: fit a model into memory without assuming unchanged quality
Understand reduced numerical precision, check runtime compatibility and compare the same tasks before replacing the original model.
huggingface.co