[hprotagonist]

I have no guarantee that I’d get into grad school at MIT even coming out of MIT’s undergraduate program, but the chances are significantly higher.

are you sure about that? Multiple tier 1 schools have a semi-official policy against reacceptance of undergrads into grad; idea being that it’s better for your development to get exposed to different university cultures.

[jwozn]

Can you provide any citations for this? I don't think it's the case for MIT

[hprotagonist]

you won't find it written down anywhere. It's part of the culture in more than a few places, though.

[Throwawayaerlei]

You'll find it "written down" in Feynman's autobiographies, he did MIT then Princeton, but that's for physics were there was much more agreement in what physics was. I was informally told the same is true for MIT's chemistry department.

Computer science has not been the subject of thousands of years of study and there's much less agreement on a great deal of it, so there's a higher chance a department will accept its own. But a MIT degree, especially if you do some research, will put you in a good position to get into a very good graduate school. And that should be no problem given what a self-starter the OP is.

ADDED: Also just notices MIT now has a "SuperUROP" program: https://www.eecs.mit.edu/academics/undergraduate-programs/op...

"The Advanced Undergraduate Research Opportunities Program, better known as SuperUROP, is designed for MIT juniors and seniors seeking an advanced research experience working closely with a faculty advisor and producing publication-worthy results. Since 2017, the yearlong program has also been open to students in the School of Engineering and the School of Humanities, Arts, & Social Sciences (SHASS)."

Back to getting into a good graduate school, the single best way to get admitted is for a professor at the other school to be told by a professor at your school who he personally knows that you can do research; success in the SuperUROP program would also help demonstrate that, especially if you can get a paper in the process of publication. The other big issue of proving general mastery of the field would be pretty well assumed if you graduated with good marks from MIT.

[hprotagonist]

> Back to getting into a good graduate school, the single best way to get admitted is for a professor at the other school to be told by a professor at your school who he personally knows that you can do research;

that is 100% the correct answer. “hey alice, it’s bob. undergraduate charlie just applied to your department; he’s in my lab now and kicking ass” will trump nearly every other consideration, if alice has funding and tenure.