Kyle Maclean

1.8K posts

Kyle Maclean banner
Kyle Maclean

Kyle Maclean

@AnalyticsKyle

London, Ontario 가입일 Kasım 2015
200 팔로잉458 팔로워
Kyle Maclean 리트윗함
Harvey Lederman
Harvey Lederman@LedermanHarvey·
Two thoughts: 1) academics are so shortsighted! We’re faced with a transformative technology and all anyone can talk about is plagiarism. Use blue books! It’s not that big a deal. 2) AI may very well destroy the university as we know it. It will change what it makes sense to teach and more importantly how it makes sense to teach. We could offer thousands of classes with AI tutors guided by faculty, we could teach new skills that help people navigate the new world, habits of mind could be more important than content, maybe more radically human society could focus more on leisure/schole and the university could take up more of our lives as we focus on fulfillment. We supposedly have imagination and creativity! Let’s start using them instead of focusing narrowly on one issue with assessment
Brian Leiter https://bsky.app/profile/brianleiter.@BrianLeiter

"AI will destroy universities" leiterreports.com/2026/04/06/ai-…

English
30
45
292
55.7K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@stoobe @IMAO_ Well, assuming those traits are random, that's actually not a major issue. But when you have correlated errors (particularly across states) that's when there's a BIG problem.
English
1
0
1
16
Stoobe
Stoobe@stoobe·
@IMAO_ Yeah, the challenge for everyone was that polling is hard: no one wants to pick up the phone or text back to strangers. There was room for prediction error and Nate acknowledged that better than almost anyone else.
English
1
0
8
918
Frank J. Fleming
Frank J. Fleming@IMAO_·
Let's not revise history -- Nate Silver got screamed at by the left for saying Trump had a 1 in 3 chance of winning when everyone else said Hillary had a 99%+ chance of winning in 2016.
Carl@HistoryBoomer

The richest man in the world doesn't understand probabilities! Tonight I'm playing poker. I'll stick $800 into a pot with top set against Vinny's flush draw and gutshot. I'll be 75% fave to win, but Vinny will spike a 9 and take my money. That doesn't make 75% "wrong." lmao

English
27
55
1.6K
78.2K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@arpitrage I suspect that without incentive in the opposite direction, the administrative incentive would be to just accept all absences. It's easier, and less risky in a sense, for them rather than arguing. We already see this with accommodations.
English
0
0
0
259
Arpit Gupta
Arpit Gupta@arpitrage·
Proposal: there should be a University-wide office whose job it is to manage student absences. If you have a good reason for being absent, you talk to them; and they inform the relevant faculty (passively through some dashboard) what students have good reasons to be gone.
English
46
14
380
89K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@mahaniok Delve, for instance! It's difficult to buy that a couple of 21 year old are super passionate about compliance. I think the more reasonable conclusion is they're super passionate about being founders/YC attached and that image.
English
0
0
0
38
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@mahaniok This is merely an observation, but the other thing I notice is young people whose sole aim to be a "founder" or get funded by YC. They don't seem to be passionate about their business so much as being able to associate themselves with these supposedly high status terms.
English
1
0
2
159
Ihar Mahaniok
Ihar Mahaniok@mahaniok·
The story of Delve (fraudsters and thieves) is just one of many examples of the rapid decline of ethics in Silicon Valley over the last two years. There are many more with recent scandals with Higgsfield, Cluely and so on. This is a logical extreme outcome of the "war on political correctness" and "anti-woke". The behavior that was not acceptable before 2024 suddenly became acceptable, and even venerated (example: a huge round by a16z in Cluely, an extremely unethical company led by a founder with zero morals). As a VC, I see it clearly in the dealflow. We always saw some founders with no integrity before. Before 2024, they flamed out quickly. But now they are not ashamed of it, but rather boasting that their lack of integrity is a competitive advantage. Now they get funded by YC, a16z and other funds, getting veneer of respectability. This needs to change, for the sake of the startup ecosystem and humanity.
Gergely Orosz@GergelyOrosz

This Delve story is going from “very bad” to “beyond very bad.” Apparently Delve’s founders were so shameless that they 1. Charged a fellow YC company (Sim) their full free for “auditing” (that turned out to be fake) 2. Then ripped off Sim’s IP, and sold it to customers for $$

English
46
21
457
65.8K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@DavidAstinWalsh "Tongue-in-cheek" in service of what though? The only thing I can think of is, essentially, "attention" for lack of a better word. At the expense of general conversational quality, i.e., I doubt productive conversation starts with someone calling for someone to be fired)
English
0
0
4
1K
David Austin Walsh
David Austin Walsh@DavidAstinWalsh·
Being slightly tongue-in-cheek here, but it is absolutely a disservice to whatever is left of the ideals of higher education to baldly admit to using AI to write and saying that you encourage your students to do the same.
English
5
2
48
20.4K
David Austin Walsh
David Austin Walsh@DavidAstinWalsh·
People like this should be fired.
Alexander Kustov@akoustov

@eyebaws1 So what? OP is not a student. Besides, I also proudly use AI to write as a professor and encourage my students to use it responsibly.

English
10
12
308
39.8K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@AJP_PhD @akoustov @eyebaws1 You're saying any AI usage is bad/wrong? What is your logic there? How can you claim that without knowing the myriad ways it is used?
English
0
0
1
23
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
I imagine a world where sending an email costs $0.01. This would add friction to large scale spamming. This may even be decided by the recipient. CEO or public figures may set a high "delivery rate".
Alex Imas@alexolegimas

I’ve been beating this drum: by making traditional signals of effort/information obsolete (eg writing a thoughtful email), AI will *increase* the value of social capital and networks. This is not a good outcome: information technology was supposed to “flatten” interactions. But I don’t see a solution on the horizon.

English
0
0
0
26
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@PlusEVAnalytics Right but I mean presumably it has a similar incentive - avoiding some sort of legislation.
English
0
0
1
25
Plus EV Analytics
Plus EV Analytics@PlusEVAnalytics·
@AnalyticsKyle It's not the same. With HHR they go out of their way to make it look like you're NOT betting on a horse race.
English
1
0
0
728
Kyle Maclean 리트윗함
atlas
atlas@creatine_cycle·
it's really heartbreaking to see the idea of "attention is all you need" go from the literal transformer to being weird and off-putting online in order to sell slop software
English
6
4
156
5.9K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@Afinetheorem Going through this now. Was an AllDayTA customer this semester, and now my university legal office seems to not be happy. (I paid out of my own pocket to pilot it!)
English
1
0
0
12
Kevin A. Bryan
Kevin A. Bryan@Afinetheorem·
Half agree but bigger problem in EdTech is that schools & univ are literally the worst customers. Technically uniformed, small contract sizes, huge bureaucracy. So a lot of EdTech vendors specialize in fighting bureaucracy, not quality. (Obv alldayta.com is exception!)
Andy Masley@AndyMasley

imo as someone who was on the ground teaching for 7 years, the terrible state of ed tech isn't because technology is somehow unable to help with learning, it's more that the specific field is flooded with snake oil bc of a few bad external dynamics.

English
2
1
17
3.3K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@anup_malani The issue is that oftentimes you must review intermediate steps to have confidence in the output. It IS possible to review the full sequence but that is still more than just the "final" output.
English
0
0
2
26
Anup Malani
Anup Malani@anup_malani·
Practical application: AI can spec, write, test, and deploy a feature. The instinct is to insert human review at every stage. But that breaks the chain and kills the gain. Better: let AI run the full sequence, then review the output. Design around chains, not checkpoints.
English
2
0
1
424
Anup Malani
Anup Malani@anup_malani·
Ask AI to handle one step of a project and you get a small boost. Ask it to handle ten steps in a row and you get a massive one. But leave one human step in the middle and the whole chain breaks. @johnjhorton and co-authors show why.
English
2
3
6
3.6K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@0Beanie05923291 @MrDanielBuck I've never understood this argument. If the degree says someone received 1.5x, there's no privacy issue. Nothing medical or health related is being disclosed.
English
0
0
0
26
beanie0597_2.0
beanie0597_2.0@0Beanie05923291·
@MrDanielBuck Colleges are entitled to that information but it’ll never happen because of privacy issues.
English
3
0
26
11.8K
Daniel Buck, “Youngest Old Man in Ed Reform”
Daniel Buck, “Youngest Old Man in Ed Reform”@MrDanielBuck·
Curious everyone's thoughts: Any student who receives accommodations receives a diploma and standardized tests scores with a flag noting that they got accommodations. A small counter-incentive to accommodation-nation, but something is better than nothing, no?
English
61
12
401
206.6K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
@jessesingal Also, don't have some standard that the majority of humans don't also meet.
English
0
0
0
49
Kyle Maclean 리트윗함
Jesse Singal
Jesse Singal@jessesingal·
NEW RULE*: You can't make a sweeping claim about how AI doesn't *really* 'think' or 'understand' or 'reason' unless you have a specific definition of the term in question in mind! Otherwise the conversation will just spiral into vapidity. *that I am politely suggesting
English
56
7
202
19.8K
Kyle Maclean
Kyle Maclean@AnalyticsKyle·
The only edit I'd say here is that rather than trying to portray it in a bad light, he was trying to tell a good story. Same result but different motivation.
Iain MacMillan@IainMacBets

He didn't "lose $10k". You gave him a free $10k bankroll and told him he's not personally responsible for any losses. He's also ultra-religious (Mormon) and knew nothing about sports betting, or gambling in general, before this "assignment". At best, this was a poor representation of the sports betting experience by the general population. At worst, this was an experiment done in bad faith to try to paint the sports betting world in a bad light.

English
0
0
0
89
Daniel Sempere Pico
Daniel Sempere Pico@dansemperepico·
You guys all run Claude Code with claude --dangerously-skip-permissions right? Because otherwise how in the world can you sit there accepting every single permission when building something?
English
475
22
2.2K
286.3K

탐색

@stoobe @IMAO_ @arpitrage @mahaniok @DavidAstinWalsh @AJP_PhD @akoustov @eyebaws1