Hacker News: acbart

New comment by acbart in "Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes"

acbart — Thu, 04 Jun 2026 05:58:57 +0000

Yep, I'm starting to hear this more and more. Matches my local data. It's a very massive and visible shift in DFW rates.

New comment by acbart in "Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes"

acbart — Thu, 04 Jun 2026 05:51:51 +0000

The fact that you are talking about Dan Garcia, a huge figure in computing education research and an excellent teacher, and the Beauty and Joy of Computing curriculum makes this hilarious. You should look up some details about both.

New comment by acbart in "Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes"

acbart — Thu, 04 Jun 2026 05:22:30 +0000

You'd be amazed at how many students we know are obviously cheating because the logs reveal that they copy pasted a long, complete answer within seconds of opening a problem for the first time, full of sophisticated code constructs that we didn't teach them, and lot's of nicely formatted comments. Sometimes they even copy/paste the entire GPT output and then format it down.

New comment by acbart in "Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes"

acbart — Thu, 04 Jun 2026 05:18:47 +0000

In the spring, but not the fall?

New comment by acbart in "Racket v9.2"

acbart — Sun, 31 May 2026 23:44:30 +0000

It's an incredible curriculum. I have a lot of fond memories myself. I frequently wonder nowadays if it isn't the right approach to force students to stop and think a bit more about "Computer Science" and not just "Programming" in an introductory context.

New comment by acbart in "Constraint Decay: The Fragility of LLM Agents in Back End Code Generation"

acbart — Sun, 24 May 2026 15:51:37 +0000

It's crazy to me that people think of Python as dynamically typed by default. Strong static typing has been an option in Python for years now, and it should just be the default.

New comment by acbart in "AI has a multiplying effect on existing technical skills"

acbart — Fri, 22 May 2026 13:56:52 +0000

Or "Age of Ultron".

New comment by acbart in "Visual Studio Code for Education roadmap update"

acbart — Mon, 18 May 2026 21:31:16 +0000

> Due to broader product alignment decisions at Microsoft, the Visual Studio Code for Education product roadmap will be coming to an end.

Visual Studio Code for Education roadmap update

acbart — Mon, 18 May 2026 21:31:16 +0000

Article URL: https://vscodeedu.com/whats-new/2645

Comments URL: https://news.ycombinator.com/item?id=48185963

Points: 2

# Comments: 3

New comment by acbart in "The AI zombification of universities"

acbart — Fri, 15 May 2026 14:18:37 +0000

Also for deadlines and social accountability. There's a reason why there was a lot less learning during the pandemic. The simple fact that most online learning advocates don't want to acknowledge is that humans learn better from other humans in person. On average, of course.

New comment by acbart in "Scores decline again for 13-year-old students in reading and mathematics"

acbart — Wed, 22 Apr 2026 21:00:51 +0000

Last time I dove into its research, I found that Math Blaster had no impact on student learning.

New comment by acbart in "Stephen's Sausage Roll remains one of the most influential puzzle games"

acbart — Wed, 22 Apr 2026 02:05:35 +0000

The witness was fine, but The Looker was much better.

New comment by acbart in "Stephen's Sausage Roll remains one of the most influential puzzle games"

acbart — Wed, 22 Apr 2026 02:04:18 +0000

Tunic is such an incredible experience. If you ever enjoyed the original Zelda and its manual, you simply must play tunic. It captures something incredible. And it has some amazing twists.

New comment by acbart in "Stanford report highlights growing disconnect between AI insiders and everyone"

acbart — Sat, 18 Apr 2026 22:07:31 +0000

Where are you seeing CS classes with increasing enrollment? Everyone I know is saying they're seeing smaller classes. Maybe some upper division from the last swell, but we're all definitely declining this year and last year, from what I'm seeing.

New comment by acbart in "College instructor turns to typewriters to curb AI-written work"

acbart — Sat, 18 Apr 2026 21:58:52 +0000

I agree, I think many people who rail against exams underestimate how important memory is to more complicated skills. How can you debug a complex application if you have to keep looking up every operator and keyword in the language you're using? It'd be like trying to interpret poetry in a foreign language but you have to look up every single noun. I'm not saying people can't do it, but it's tedious, slow, and you probably wouldn't think of them as a "professional worth paying for their service". Some amount of memorization is key.

New comment by acbart in "College instructor turns to typewriters to curb AI-written work"

acbart — Sat, 18 Apr 2026 21:56:02 +0000

Exams happen all the time in real life. Or rather, situations where you can't just look up fundamental knowledge. Job interviews, presentations, even mundane work tasks - all these require you to know the basics quickly "The basics" are relative, of course, but I often point out to my students: "you don't care if your doctor needs to look up the specific interactions of your various meds. You do care if you see them googling 'what is an appendix'." Proctored, in-person exams are the only reliable mechanism we have for ascertaining if a specific individual has mastered key fundamentals and can answer relevant questions about them in a relatively timely fashion. Everything else is details and thresholds - how fast do you need to be able to recall, how deep, what details are fundamental. From there, I think it's fine to hate poorly made exams, and it's a given that many folks making exams have no idea what they're doing (or don't have the resources to do it right). But the premise of an exam is not completely divorced from reality.

New comment by acbart in "College instructor turns to typewriters to curb AI-written work"

acbart — Sat, 18 Apr 2026 21:49:04 +0000

So at 50%, someone who uses AI to get 100% of the homework grade will earn a D (sometimes passing) if they can get at least a 20% on your quizzes, and a C (always passing) if they get at least a 40%. Did you make your exam so difficult that students who truly didn't learn the material earn less than 20-40%? Because if it was, say, multiple choice questions with four possible answers, then you can expect them to earn at least 25% just by chance.

New comment by acbart in "Agentic Software Engineering Book"

acbart — Mon, 23 Feb 2026 05:34:57 +0000

https://scholar.google.com/citations?user=hpxl9PEAAAAJ&hl=en

The researcher seems to be real, at least? Perhaps the quote has not previously been written down?

New comment by acbart in "How to effectively write quality code with AI"

acbart — Sat, 07 Feb 2026 13:55:16 +0000

It makes it easier to make sure it runs right. Code that is easier to make sure is quality code. Code that is hard to make sure is not quality code.

New comment by acbart in "After two years of vibecoding, I'm back to writing by hand"

acbart — Mon, 26 Jan 2026 23:36:20 +0000

Yes, exactly. I'm having a frustrating time reminding senior teachers of this, people with authority who should really know better. There seems to be some delusion that this technology will somehow change how people learn in a fundamental way.