we need to talk about brilliant.org

Summary

The video is a critical but ultimately qualified review of Brilliant. The narrator argues that the platform has genuine educational strengths, but inconsistent lessons, poor navigation, adult-unfriendly design, questionable AI claims, and limited support for independent projects make it difficult to recommend as it currently stands. The video uses satire and an exaggerated story about a learner named Jane, but the narrator says the underlying criticisms are serious.

Lesson quality is inconsistent

• Some newer Brilliant courses—especially in math, programming, and AI—offer interactive problems, useful feedback, and opportunities to apply higher-level thinking. The narrator considers many of these lessons pedagogically strong.
• Other, less-updated lessons are mediocre or poorly aligned with Brilliant’s stated approach. The opening example is a quantum-computing lesson that begins with several pages of explanation before offering a problem, rather than putting problem-solving at the center.
• The narrator’s conclusion is not that Brilliant’s marketing is wholly false, but that only some of its lessons live up to the advertised standard. Poorer lessons should be improved or archived.

The platform can be a poor fit for adult learners

The narrator discusses Malcolm Knowles’ model of adult learning, which identifies six common characteristics:

• Need to know: Adults want to understand why a subject matters.
• Experience: They bring prior knowledge that learning should build on.
• Self-concept: They want some control over their learning decisions.
• Readiness: They are drawn to learning with a clear purpose.
• Orientation: They tend to prefer learning organized around solving problems.
• Motivation: Internal factors, such as personal satisfaction, are important.

The fictional example of Jane illustrates how Brilliant may frustrate an adult who already knows some material:

• Navigation can fail to find the topic the learner wants.
• Lessons may make learners repeat basic material before reaching the relevant content.
• The platform may not clearly explain what a lesson covers.
• Free users cannot freely skip ahead, while premium access can remove some restrictions.

The narrator sees these as more than inconveniences: they can waste learners’ time and make it hard for adults to direct their own education. He contrasts this with Duolingo’s ability to assess prior knowledge and let users skip ahead.

AI has potential, but Brilliant’s claims and transparency need work

• The narrator does not reject AI tutoring in principle. He sees potential value in well-designed tutoring tools, particularly where teachers have limited time or students need individualized support.
• He describes Brilliant’s Koji as a work in progress. Its ability to interact with lesson content may be useful, but in his experience, it sometimes highlights irrelevant material and does not understand a student’s strengths and weaknesses as a good human tutor would.
• He objects to claims that Koji is “the best tutor you will ever have,” arguing that the evidence and current performance do not justify that confidence. He also says interactive visual teaching does not necessarily require AI, pointing to Desmos as an example.
• The narrator says Brilliant’s privacy policy permits sharing or selling user data to third-party services unless users opt out, and criticizes the opt-out process as difficult to find. He also wants more transparency about how the company uses AI and what human review is involved.
• He reports adapting an AI-tutor evaluation framework with four dimensions—feedback quality, learning effectiveness, interaction experience, and user trust—and scoring Brilliant about 33 out of 64, while acknowledging that his assessment was generous.

Brilliant may not adequately support independent skill-building

The narrator tested the platform by learning Python and attempting an independent project.

• He found the course useful for learning to read code and appreciated its structure, but says it did not teach him how to build a project from start to finish.
• In particular, he says the course did not cover Python’s input() function, which he needed for a simple interactive program.
• Brilliant did not provide a Python console, and the guidance he received for working outside the lessons was inadequate. He describes the course’s focus on keeping learners on-task as obstructive when he tried to ask relevant questions or explore beyond the prescribed material.
• He later used Khan Academy and found it offered tools and direction that better supported continued practice.
• His broader criticism is that a platform can promote independent projects without adequately teaching learners how to undertake them. He believes Brilliant should make its programming pathways and recommended external tools clearer.

The company’s history and its capacity to improve

• The narrator recalls Brilliant’s earlier identity as a community-oriented problem-solving site, where users shared and discussed solutions. He regrets that this community role has diminished.
• He cautions that it would be misleading to turn that change into a simple story of a once-great company losing its soul. That narrative would be compelling, but it would not fairly reflect his experience.
• After contacting Brilliant, he was eventually connected with Alice and then Blake Farro, the company’s chief content officer. He says Blake checked one of the problems the narrator had raised, confirmed it, and fixed it without repeated prompting.
• The narrator views this responsiveness as evidence that Brilliant is willing to improve when problems are brought to its attention. He says he has communicated with the company about the platform’s shortcomings and hopes it will act on them.

Changes the narrator asks Brilliant to make

1. Support independent creation: Publish a clear resources page explaining how learners can use the Python course to build projects, including what tools and outside resources they will need.
2. Offer a non-AI subscription option: Let users pay for Brilliant’s human-created lessons without having to pay for AI features or tokens they may not want.
3. Archive or revise weak lessons: Remove outdated material that does not meet the company’s stated educational standards.
4. Let free users skip lessons: Provide a way to move past material a learner already knows, ideally using an assessment to check readiness.

Overall conclusion

The narrator does not recommend Brilliant at present, despite believing that parts of the platform are genuinely good and that the company has the potential to improve. He argues that the platform’s strengths are undermined by uneven content, restrictive navigation, adult-learning barriers, limited support for practical projects, and overconfident or insufficiently transparent AI messaging. His final assessment remains open-ended: Brilliant may become recommendable if it addresses these issues.

Speakers and sources featured or referenced

• KatzVE / the narrator: The video’s reviewer and primary speaker.
• Sue: A Brilliant representative or founder quoted discussing the company’s AI and content-review practices.
• Alice: A Brilliant staff member who connected the narrator with the content officer.
• Blake Farro: Brilliant’s chief content officer, whom the narrator says he spoke with privately.
• Koji: Brilliant’s AI navigation and tutoring product, discussed and demonstrated; not a human speaker.
• Malcolm Knowles: Cited for the six characteristics of adult learning.
• A 2011 article on active learning: Quoted on the difference between superficial activities and learning grounded in constructivist principles; the article’s author is not identified in the subtitles.
• An article or framework for evaluating AI tutors: Referenced for its four evaluation dimensions; the source is not named in the subtitles.
• Brilliant’s marketing materials and privacy policy: Discussed as sources for the company’s claims and data practices.
• 3Blue1Brown, Khan Academy, Duolingo, and Desmos: Referenced as examples or points of comparison.
• The narrator’s friends: Asked to assess his beginner Python project.
• Reddit users: Mentioned as a source of comments about why people may not recommend Brilliant.