AI
I am sometimes surprised that some students imagine that they no longer need to learn skills because AI can give the answers. There is some notion that they can build their future on top of the AI chain.
If AI really could or can eventually answer the questions, then what will they be useful for? Why would companies hire expensive employees whose skill is asking ChatGPT? And more prosaically, how would they ever know whether "their" answers are right or wrong if they have no understanding of how they came about? Yes, the answers may work in a course exam, but how will they explain "their" answers in a meeting where their peers are asking them how they came to their conclusions?
ChatGPT and its siblings serve a useful but limited role. They are indeed good for quick "starters" getting over the initial hump of many learning curves. Large language models are, in many ways, good alternatives to search engines. If others have solved a specific problem already, LLMs may find similar string configurations in their internal lookup tables. If not, AI probabilistically picks the next word, which may well be hallucinating and wrong. Today's LLMs seem very good at writing convincing answers, not correct answers.
If I need a quick skeleton for a computer program to get started coding, especially in a computer language that I have never or only rarely used, it's good. It has pieced together many starter programs from the web, and picks the most common version. However, it cannot implement deep logic.
I have tried this many times on many problems. What AI produces is mostly garbage — sometimes convincing but rarely correct. Here is my analogy. If you ask AI to create a program to calculate square roots, it may just create a lookup table. When you try the program, it will look as if it's working: you put in 2, and the answer is 1.414; you put in 4, and the answer is 2. But it really isn't a problem-solving program. There is no deep understanding of the concept. The more complex your needs are, the more obvious the shortcomings will be to someone who really understands the subject. Do you really still need to understand what square roots are? Isn't it enough that you can ask AI, or google, or a calculator for the answer? Yes and no. Calculators are one-task devices, guaranteed correct. Google has square roots hard-coded, so it is correct, too. AI is most likely correct, too. But if you don't know what your AI program is supposed to be solving and you don't have the conceptual understanding and ability to solve at least simple versions of the problem yourself, then how would you ever know whether the AI answer was garbage or not? You couldn't even check it. You would be like the boss to a bunch of research assistants who are on LSD. They specialize in sounding convincing, but they are completely untrustworthy. As soon as you have a problem that exceeds the basic capabilities, AI can suddenly and unexpectedly (and without warning) give wrong answers. You don't need to be an expert in square roots, but you do need to understand the concept behind square roots if you want to use them. You need to be able to check the answers.
The fact that AI can get the most basic versions of many questions right — which means that it is especially good on exam questions which are the kind often posted somewhere on the Internet — doesn't give you the understanding you need. Without the conceptual understanding inside the gray matter in your skull, you will never be able to reason up to real-world problems. You would be building a house without a foundation in an earthquake zone.
It is the talent of true general intelligence, hopefully yours, to reason and build up from a solid understanding of the basics to answer new and different problems. ShitGPT cannot do this, and if you think you don't need to know the basics because ShitGPT can look it up, you will not be able to build up, either. You will be useless. You won't be able to solve the real-world problems that you will encounter (at least with any sort of reliability and confidence). If you let ShitGPT answer your quiz and exam questions, you are cheating yourself and not the instructor. You may get good grades, but your (lack of) education will be useless. Save yourself the tuition.
This is why we still teach the basics. We don't teach them for their own sakes. We don't expect you to encounter Finance 101 questions in your future career. We teach it because you cannot reason up to real-world problems without understanding of the basics. Your Education would be a house without foundations!
2025- Teaching
408:
EMBA Hybrid Monthly
457C-02:
ASAM
404:
MSBA 4-Week Finance Module
Corporate Finance
corporate finance
A free online textbook with a low-cost printed version, full instructor materials, a course administration site, a student self-test site, and much more. The best book there is, regardless of price. I also posted a presentation to the FMA at here
course and test site
My course pages are now hosted on syllabus.space, which is much simpler for both instructors and students than any alternatives. (Students not in a course using the system can also take free finance quizzes here.)
MFE 404
A first-to-second Corporate Finance and Economics course for MFE students, with emphasis on financial statements, comparables, and data and computer programming.
- The 2021 syllabus is syllabus-2021.html.
- Classnotes are on the course website and also here.
Older Courses
Informational Cascades, June 2023:
Stockholm Mini PhD Class
- see learn more.
Stanford GSBGEN 341.1 Winter 2023:
Climate Change, Economics, Technology
- Course Schedule: BC104 Classroom, Tue/Fri 14:50h-16:10h. 10 Weeks.
- Stanford site.
- Preliminary Syllabus
- ... based on Textbook: Global Climate Change.
298E-04:
Energy, Climate Change, and Finance
- Course Schedule: Wednesday, 19:10-22:00h. 10 Weeks.
- Climate Change Syllabus
- ... based on Textbook: Global Climate Change.
254-01 (PhD):
Empirical Corporate Finance
- Syllabus (not offered in 2022-2023)
Other
arrogant?
I have read many times in my teaching reviews that some students consider me arrogant. I don't think I am, but I do challenge my students the same way I challenge my colleagues and the same way I expect to be challenged myself by others. This is academia and my mannerism is just this way. Perhaps the best way to describe some students' misperceptions is by analogy with this famous scene from Lawrence of Arabia.
other
Helping less privileged.
A free IPO teaching case
Old but not badly aged
No answers
advice for
Students Undergrads MBAs MFEs PhD Applicants and PhD Students Profs finintro
Feel free to ignore. It is worth what you are paying for it.