jventura
What's happening with Namecheap? I've been a user for a long time and haven't noticed anything.. Maybe I'm one of the frogs being boiled!
But yeah, a blog post is not a short message. This one here is a short message and I didn't use any LLM. But I can guess that people that are not so fluent in english as myself may need it..
Is there any other mechanism replacing it? As far as I understood, while following the LLMs from Scratch book, the dropout is a mechanism to prevent overfitting, and it makes a lot of sense (basically cripple some neuron connections randomly during training such that their weights aren't updated).
Edit: well, I could have searched on the internet, but other people may stumble on your post and you seem quite knowledgeable of the inner workings of these things. :)
Basically, most feminine words end with an “a” while masculine words end with an “o”, although there are some words that end with other letters whose genders is not that easy to identify. For instance “árvore” (tree) ends with an “e” and is feminine (“a árvore”), while “peixe” (fish) also ends with an “e” but is masculine (“o peixe“).
By the way, there is also another "infrastructure" element that is used as art in Portugal - calçada portuguesa [1], which is also very interesting.
Of course that the "succesfull" groups implemented their solutions, tweaked them and, tried their best to escape from race conditions and were able to talk about the performance/memory/etc. tradeoffs. We have a grid to grade for those and other items. They did not outsource all of their thinking to an LLM.. That is what I mean as taking responsability for their code, maybe not the correct word, but the one that came to my head when I wrote it..
How could we "force" the students to use an LLM that confronted their doubts with more questions? We could tell them to start each chat with a specific prompt (to use the socratic method, etc), but they could eventually jail-break it..
But nevertheless, I like your idea! This is something that a document highlighting methodologies for students on how to use LLMs effectively could/should contain..
I have colleagues that are teaching for more than 30 years, few years away from retirement, who suddenly have been confronted with a new way of doing things. Those are the ones that are still insisting on doing practical projects, etc. I've only been doing this for 20 years, and I'm quite lazy (worked previously as software engineer), so I've moved to those practical tests. I guess that there should probably exist a class or workshop to teach these students how to use LLMs effectively, but as I said, this technology and its implications is quite new.
Personally, what I did was to give them the "lecture" in the line of that they do not understand what the machine has generated, that is not the way a true engineer does, try to do some parallel with things like an LLM designing a bridge and civil engineers building that bridge, and a fatal flaw collapsing the all thing, etc.
In other words, we do not have a formal system in place, it's all talking and convincing them. Obviously it's a big enough problem that should deserve more investment in solutions, but we are all overwhelmed by other tasks. Maybe LLM studios should be held responsible for all these "disruptions" and provide solutions to problems they created! :)
Yes, if they are "responsible" for the code delivered, where responsible means they understand the code, the architecture, the decisions made, etc.
In this case, the students had to invent multiple strategies to solve a specific problem. The "successful" groups did a mix of generated and hand-crafted code (don't know percentages), implemented different strategies and knew their plus and minuses, could change the code in a timely manner to accommodate some of my requests, etc. The "unsuccessful" group couldn't do any of that.
I'm not anti-AI (and really, what could I do if I were?) since I use it myself, I'm just anti-slop, especially from my students.
But in reality I've been slowly transitioning from group projects (for a subset of the grade) to "practical tests", where they must implement a significant subset of a larger project in a 2h class. Still experimenting though.
To fellow professors, when you're suspicious my suggestion is to appeal to their honesty (like "let's be honest, how much of this code is yours, and how much is ChatGPT's?") and offer some empathy and understanding (like understanding they may had multiple deadlines in the same week, etc.). Nevertheless, don't miss the chance to give them the lesson on how is the correct way of doing things. The way to catch these students is to find the same signs of yesteryear copying from other students (which in essence is what copying from an LLM is, although the number has increased because they found us professors unprepared for the volume).
The other two groups also used LLM but in a high-level and architectural way. They were clearly responsible for the code (even if they didn't wrote it 100% manually) and could explain their reasoning and strategies used to solve the problems.
Me and my colleagues still have a lot of projects to review, and I asked them to keep the score of the number of projects like these, but so far, the score is 1 in 3 (33%).
The sad state we got ourselves into..
Oh, and fahrenheit, what the hell it means? 0ºC means ice, 100ºC means boiling water, 40º feels summer around here..
I guess I'm saying that you understand the values of the imperial system because you're used to them, as I'm used to values in the metric system..
Personally, I've used XP a lot back in the day, but don't remember much of the details like most users are reporting here, so I really liked to play with your website, and would definitely hire you if I was in such position.
Good luck!
I have a python astrology library [1] which I use to generate my astrology chart, as text, to feed it into an LLM (Gemini 2.5 Pro). I then ask it questions about personal characteristics (the easy parts), but, more interesting, I can also ask it to consider some scenarios and it answers back with several hypothesis and how well it fits my character, personal goals, etc. It's like talking to a friend that knows you very very well.
Lately I've been working with a technique called Primary Directions. It's a predictive technique that tries to describe events in your life by means of astrological symbols (things like "Opposition of Saturn reaches the MC by 34.5 years old", which means something "bad" for your career or social position) and use it to check if a specific scenario has worked for me previously and will work in the future, and to ask it for other scenarios that match my personal characteristics and predicted symbolic events. I find LLMs, specifically Gemini Pro, quite good at these kind of things.
I also have fun "playing" with other peoples charts. For instance last night I gave it my chart and list of primary directions to Gemini and asked it if it could find who it was. It said Kurt Cobain. Quite off! But it described a lot of events that could fit my primary directions, like for instance, that Kurt Cobain got his guitar at his 14's or the one at 27 where he died. I didn't die at 27 (but got very ill) and also got my first guitar at 14. I'm also a musician, although an amateur one.
If you're into these kind of things, I created a gist [2] that you can feed into an LLM and talk to Kurt Cobain's chart. Note that it doesn't mention anywhere that it's Kurt Cobain. For fun, ask it something like "Considering the chart and the events predicted by the primary directions, in which ages will this person have some success of public visibility". In my case it answered, among others, 13-14 yo something related to "success, popularity, academic, sports or artistic achievement" (Kurt Cobain seems to got his first guitar at 14, and discovered his vocation), and 23-24 "beginning of career, marriage, or first step that puts the person on the 'map'" (release of the Nevermind album that catapulted Nirvana to the world stage). You can then ask it to match the events to Kurt Cobain, and it will find the real life events that seem to match it quite well.
I find that LLMs are quite good at generating hypothesis, multiple scenarios, and I'm still exploring their strengths and weaknesses (and of astrology as well).
[1] https://github.com/flatangle/flatlib/
[2] https://gist.github.com/joaoventura/68e0aed7c49c389347df98ec...
Maybe the idea can help you starting things?
It also helps that, sometimes, when the tasks are big, I convince myself that I can finish it later. Many times I do not have to finish it later..
All gas stations closed because they could not sell gasoline/diesel. Today there are lines on all gas stations, people filling their car tanks and bottles..
Oh, let me tell you about electric cars! Many people had to spend the night somewhere away from home because they could not charge their cars.. My sister (with her job's electric car) had to stay the night some 200km away from home, and since the ATMs (Multibanco) didn't work, she didn't have physical money to pay for food. Luckily a stranger paid for the food (yogurt and some cookies). Petrol cars, because of their range, had better luck!
Pure fear and panic..
I can only blame the authorities (Portuguese/European) for not having contingency plans for keeping people informed, and thus letting fear spread like wildfire.
Not here where I leave (not USA)..
> If that means no more coursework, fine, scrap the coursework or require it to be done under supervision as well.
I have a course on building web apps. Not 2h prototypes, but apps that take days to build. Do you think it's the same as a 2h exam?
> CS departments especially have a wealth of options available to them via automation. Record screens on systems without network access, require students to be patted down at the door to stop them bringing in hidden phones, and watch them carefully as they work both in real time and do spot checks on the screen recordings afterwards. Or for that matter, use AI to do it.
Do you think we are police officers or what?! There's a limit for what we are able to do, and what students are able to tolerate..
> The alternative is to just see universities be bulk defunded in future as a failed experiment: see what the Trump admin is doing (...) If degrees are worthless because universities won't do what it takes to stop cheating then what's the argument for preserving student loans next time there's a debt crisis?
Ok, I get your point now! You probably live in some third world country where there's lots of wealth inequality and the state does not help its own citizens. Here in Portugal (and I guess almost all of EU) the tuition is very cheap (around 700€/year), so I wouldn't consider failed experiments for now..
Not all assessments are exams, eg: projects, and students cheat even in exams with professors present, eg: the OOP exam I mentioned before, where we had 2 teachers for 40 students.
Maybe you’re suggesting 1 professor watching 1 student on a 2,5h exam? For 40 students we would need 40 professors.. We don’t have that number of professors in out departament..
This semester I'm teaching a web development course (fullstack development), and my policy is that the project must be done on github classroom repositories, and I'll be asking clarification (face to face) on some of the commits. They can use whatever they want (stackoverflow, chatgpt, whatever), but they better know how to explain their commits to me. I don't know if my approach will work in the end, but I surely got their attention.
I'm doing this because it got so bad that, last semester, on my Object Oriented Programming course, even using moodle with the Safe Exam Browser on and an instance of Visual Studio code to try the code, we caught lots of cheaters. How they were doing it? By installing co-pilot plugin. How did we caught them? Some students solve all the exercises in 10 minutes, others had comments that were clearly made by AI, etc. etc. Of some 15 students we caught, only 3 came to us to review the exam.
Hard problem to solve..
Been there many times, but I found that it’s best to aim for good enough, given the resources and constraints, than 100% all the time. I’m a recovering perfectionist, so it’s not always easy to reach that balance, but I try it anyways. When I do not, my anxiety kicks in, and I’m forced to slow down.. I use to be able to ignore it (the anxiety) and keep pushing, but not anymore..