The context of this module is for students pursuing a computer science degree, possibly in their freshman or sophomore year.
There are many factors that impact the CS reality in 2026.
For most freshman and sophomore students, the COVID pandemic impact is significant. In order for the youngest student to be a freshman or sophomore in 2026, the student would have been in eighth or ninth grade or so. This is significant because eighth and ninth grades introduce important concepts in math that are foundational to computer science.
The impact of the COVID/pandemic is not limited to the depth of knowledge due to the modality of delivery. The change to the pass/no pass scale and the consequent “pass-along” effect changed the math readiness of high school graduates.
Furthermore, there is also the attention span muscle apathy issue. During the pandemic, many high school classes were online and asynchronous. As a result, students did not get the opportunity to develop their attention span (brain) muscles.
Where classes are online and asynchronous, most students do not learn to take notes. The lack of note-taking skills severely impacts college-level success.
Essentially, a GPT can complete almost all community-college-level homework assignments that require writing, including coding.
LLMs have an attention/context window. This window is measured by the number of tokens. A token is roughly a word or a punctuation mark. The context windows of the most capable LLMs are measured in 100,000’s of tokens. While this number is huge by human standards, it is actually not that much. This is because the context window includes all the text in a conversation, including the content generated by the LLM.
The human mind does not operate by raw text. Cognitive science is starting to understand how the human mind works. It suffices to say, however, that the actual context that can be active for a human can easily exceed that of the context window of an LLM. The caveat is that this is not possible without training. This is to say that the physiology of a human allows most people to run a Marathon. However, for an individual to run a Marathon, a lot of training is needed.
The short-term memory of a human is limited to 5 to 9 items, and the retention time is measured in a double-digit number of seconds. A phone number with 10 digits is about the limit of the short-term memory of most people.
Then why do many students feel that they can retain all the material while in a lecture? This is because a professor often repeats the same concept. Each mention of a concept is a refresh of the memory, extending the retention of the concept by another 20 to 40 seconds. Some short-term memory content can be committed to long-term memory during a lecture, but that depends on many factors.
Taking notes either by writing or typing externalizes short-term memory to a form that persists until there are sufficient resources to commit the content to long-term memory. The act of taking notes already helps commit short-term memory content to long-term memory. There are multiple mechanisms at play:
The brutal answer is “you will not get hired”. And yes, that is assuming with a B.Sc. in Computer Science. As of the beginning of 2026, CEOs of large tech companies are vibe coding to prototype ideas. The trend is to “cut out the middleperson (software engineer)”.
Some people may be curious about “what difference does it make?” After all, if vibe coding replaces all software engineers, then whether a computer science student uses GPT to do all the homework should not matter, because there are no jobs. If there are no jobs, then selecting computer science as a major is completely moot.
The key point is that vibe coding is only productively applicable in certain cases. The “vibe” of vibe coding implies a casual use of normal language, as opposed to a programming or formal language. Natural languages are ambiguous and not suitable for specifying something (like a computer program) that requires precision. Computer science may be a relatively new field of study, but mathematics is not. The reason why mathematics has its own symbols is not only for conciseness, but also for precision. The use of parentheses, for example, explicitly states the nested nature of a structure. Most natural languages fall apart quickly when a complex concept that requires precise nesting specification needs to be explained.
As a result, vibe coding is intrinsically limited to applications where logical precision is not of great importance.