AI Fluency
Where models are dependable, where they are not, and how to tell which one you are looking at.
- What a model is actually doing when it answers
- Why confident and correct are not the same thing
- Checking work you cannot verify yourself
Research, papers and AI news
Go past the summary. See where the claim came from.
Explore AI, software and security research without losing the original work. Every entry links to a paper, article, CVE or official source and is graded Beginner to Advanced, so you choose how deep to go.
- 1 78 sources, read daily Official feeds only, never scraped
- 2 Verified, filed under 15 topics Every claim traced to its source
- 3 You search, read and follow the source Summaries are the doorway, not the evidence
- 4 Turn the lesson into practice Prompt packs show the method
- 5 Keep learning as AI moves Useful news without the noise
What are you trying to do?
Pick one. The library narrows to the topics and depth that fit.
Or filter manually
Loading the latest research...
Free prompt packs
Start with the free packs
A few prompt packs are open to everyone, no account required. Each has a read-only audit prompt and a hardening prompt with verification steps. Run one against the feature you are shipping this week.
Full access
This is the free selection. The full library is the product.
Every article selected as free is available above, with no account needed. A free account opens the broader current research feed, up to 25 entries. Basic, for the price of a coffee a week, opens the whole library, search that cites its sources, and every audit and hardening pack: the fastest way to secure the AI feature you are shipping. Premium adds prompts written for your exact situation and instant alerts. Expert adds an hour with us every month.
Every paid plan starts with 3 days free. Nothing is charged until day four, cancel before then and pay nothing. Pay yearly and two months are on us.
Start free, upgrade when you are ready Already have one? Sign in
Coverage
Building, prompting and security at three depths
Security
Building
Every entry links back to where it came from. For articles found through search, Korpalis reads the public page to classify it and write an original summary; we do not republish or store the article itself. If you publish research or news and would like it excluded from this database, email privacy@korpalis.com and we will remove it.
Learn
01
02
03
The groundwork underneath the entries
This library tracks what is happening. It reads a lot better once you know how models behave and how good prompting is structured. Anthropic publishes its course material free at Claude Academy, and it is one of the clearest starting points we know of. No affiliation and no referral: just the reading list we would give you anyway.
Build with Claude
The developer track: the API and Claude Code first, then agents that call real tools.
- Claude Code and the Claude API from the ground up
- Agents built on MCP, skills and tool use
- Written guides as well as video lessons
Claude for work
For rolling it out to people who will never open a terminal, and keeping that sane.
- The everyday tools, for non-engineers
- Handing off whole tasks with Cowork
- Setting an organisation up properly
More from the same source: structured courses, written tutorials and worked use cases. All of it is free, and none of it needs a Korpalis account.