The teen AI programme market has filled up with expensive summer offerings whose main product is a line on a college application, so it is worth saying that AI Camp is one of the more substantive options in it. The bootcamp runs three weeks, five days a week, about four hours a day, for roughly sixty hours in total, entirely online and open to students from thirteen upward. The defining feature is class size. Four to six students per instructor is genuinely small, and unlike a lot of programmes that advertise personal attention and then run cohorts of thirty, this one appears to hold to it.
For a teenager who has never written code, having someone who can actually see their screen when they are stuck is the difference between finishing and quietly giving up. The structure is the other thing it gets right. The three weeks are organised around building and launching one AI product rather than working through a syllabus of concepts. Students come out with something deployed that they can show a person and explain.
That is the correct instinct for this age group and this subject. Interest survives when there is something to point at, and it dies when the whole experience was a sequence of notebooks that produced nothing. The zero coding prerequisite is also handled honestly rather than being a marketing line, with the early days spent making sure everyone can actually operate the tools. Now the parts I am less comfortable with.
The price is around $2,349 for sixty hours, which comes out near forty dollars an hour. Small group instruction costs money and I am not pretending it should be free, but the comparison set matters. A teenager who is genuinely self driven can work through Kaggle Learn, Karpathy's videos and a free Colab account and get further technically for nothing. What you are buying here is structure, accountability and a peer group, so the question is whether your particular child needs those things.
Some do, decisively. Some do not. The instructor framing is my other reservation. The site leans hard on Stanford, CMU and MIT, but in practice a great deal of teaching in this market is done by current undergraduates and recent graduates from those institutions.
That is not a bad thing in itself, and enthusiastic near peers are often better with teenagers than senior engineers are, but the branding implies faculty and the reality is closer to a well organised tutoring operation. Parents should price it as tutoring rather than as a university experience. I would also push back on the admissions argument that hovers over this whole category. Admissions officers have seen a great many paid summer AI programmes by now and they discount them accordingly.
What still carries weight is the project itself, particularly if the student can talk about the decisions behind it and what broke. AI Camp at least produces that artefact, which is more than many competitors manage, but the credential itself is close to worthless and the marketing should not suggest otherwise. Three weeks is also simply not long enough to develop depth. Students finish able to assemble an application on top of existing models, which is a real and useful skill, but not with any grounding in the mathematics or engineering underneath.
Treat it as an ignition source, not an education. My 3.7 reflects a well run programme with genuine small group teaching and the right project first instinct, held back by a price that is hard to justify against free alternatives and by branding that flatters the staffing. The sensible route is the $99 AI Foundations course first. If your teenager finishes it and immediately wants to build something else, the bootcamp is a reasonable next step.
If they do not, you have saved yourself over two thousand dollars.