dynomight
@dynomight7
Followers
4K
Following
1K
Media
522
Statuses
2K
acetone👍aspartame👍alcohol👎correlations👎ethylene👍gratitude👎humidifiers👎lists👍mbti👍p2p meth👍potatoes👍pm2.5👎seed oil👍sloths👍tubes👍ubi👎wirecutter👎
Joined May 2020
Upon reflection, I think it's fair to say Crémieux plagarized this thread from me. Try this in school: Take the whole argument and huge sections of text (unattributed) from some other source, but put some tiny offhand reference at the end. Your teacher will call that plagarism.
43
49
1K
@AgnesCallard no objection to the parenting but feel you're breaking an unspoken covenant with all the people that paid money to buy candy to give to your kids.
7
2
297
@cremieuxrecueil And do you regularly take large sections of text from those sources without attribution? You believe this is normal and acceptable behavior?.
1
0
326
Interested in thoughts on this. This basically takes the entirety of a blogpost I wrote in 2023 and turns it into a thread. I'm mentioned at the end but still. am I wrong to expect more credit?.
25
11
314
Not mentioning this was a mistake
10
7
283
If you have any doubts, I ask you to compare that thread side-by side to my original post:. I've asked several people to do this, and they were shocked. It's not a "source"—it's the whole thing. It even steals my joke about Ajinomoto and MSG.
3
0
269
This graph when around in 2018, supposedly showing that in countries with more gender equality, fewer women choose STEM careers:
2
10
191
I offered Crémieux the opportunity to choose a neutral third party to arbitrate. Since that was declined, I nominate @tracewoodgrains
5
1
188
@cremieuxrecueil How exactly do you define the word plagarism?. If I took words you wrote, and re-posted them as I had written them, would that be plagarism?.
2
0
154
@stanfordNYC It's telling that Crem is aware of these allegations, but instead of apologizing, just says "what's the problem?".
2
0
136
There's an amusing irony in the fact that if I link to my post the algorithm will suppress it. But I mean, compare these.
2
0
124
How I program now:. 1 Write spec.2 Ask AI for ambiguities.3 Fix ambiguities.3 Ask AI for tests.4 Ask AI for code.5 Shriek at 500-line functions with deeply nested branches.6 Ask AI to fix code to pass tests.7 Ask AI to fix new bugs just introduced.8 Give up, write the code myself.
2
1
114
@cremieuxrecueil Right. So then we're agreed: What you did is not paraphrasing, but plagarism.
There's an amusing irony in the fact that if I link to my post the algorithm will suppress it. But I mean, compare these.
3
0
106
Why is it so hard to imagine distant future post-scarcity utopias that actually sound appealing?. Lack of imagination? Meaning comes from struggle?.
43
5
101
If you'd like to know what Crémieux has to say about this, this is the only reply I'm aware of.
4
0
93
@cremieuxrecueil Would you like to propose a neutral party to adjudicate this dispute? If they rule that this is not plagarism, I will withdraw the allegation and apologize. Will you commit to admitting plagarism and apologizing if they find against you?.
2
2
92
@cremieuxrecueil > I don't need help to know what's plagiarism and what's paraphrasing. Please consider that this may be exactly what you need.
1
0
89
I recommend this article, in which I teach information theory+molecular biology from scratch, argue that existing definitions don't capture the intuitive idea of "information", propose a new definition that does, and argue it lies outside the current human knowledge frontier.
6
7
85
With the debate gas stoves, wouldn't it be nice if someone had estimated how much having a gas stove impacts life expectancy?. And wouldn't it be. umm. impossible to have a meaningful conversation about tradeoffs without knowing what that number is?.
6
4
72
Lots of people are (pleasantly 😀) disagreeing with this, but note that I wrote it as advice for bloggers interacting with readers, where there's a stronger norm towards politeness. I'm interested to hear from other bloggers if they agree it's empirically true.
I try to be relentlessly reasonable with people but I could probably do with internalizing this wisdom a bit more so that I don't waste my time
4
3
69
@cHHillee 👍 All the comments saying "this is just {vmap, xarray, namedtensor}" are wrong (and explained in the post), but this one is correct! torchdim is the closest precedent I'm (now) aware of.
0
0
61
For the last year or so, I've been trying to understand who is required to get permission before doing research with human subjects in the US.
3
4
52
here i try to convince myself things aren't as bad as they seem.
0
2
42
I took 13 different historical technologies and tried to imagine what would happen if large language models played out the same way. 🧶.
3
5
47
@stanfordNYC @cremieuxrecueil Crémieux claims the huge number of unattributed screenshots are OK because they're screenshots. Which. you can decide for yourself. But it's not just that—the "regular text" in the thread is also largely stolen.
0
0
47
@cremieuxrecueil Can you give any examples of activities that would constitute plagarism, according to your definition?.
1
0
46
@norvid_studies Seems likely. I know this behavior is common with engagement bait twitter hucksters, but Crémieux presents as a "serious" "intellectual".
3
0
43
@CRPBN99 @cremieuxrecueil Listing a "source" does not entitle one to re-post the words in that source as if one had written them.
1
0
42
With abortion in flux in the US, I realized I didn’t have a clear picture of how things looked in the rest of the rich world. So I researched 27 countries and tried to visualize everything in a single graphic. Mostly, I learned that things are really complicated. (thread)
1
8
34
One view:. The biggest thing we've learned is that intelligence is just way easier than expected. I don't understand how you can't find this at least a *little* terrifying.
5
0
37
@Wolframtheman Just re-using the ideas would be rude, but not plagarism. Crémieux took many large sections of text without attribution. If you haven't actually compared them—and I'm guessing you haven't—I encourage you to do so.
1
0
38
What does it mean when people say "religiosity is 40%" heritable?. I used to think we had some points from genes/environment:. Genes ★★★☆ (3/4).Environment ★★☆☆☆☆ (2/6).Total ★★★★★☆☆☆☆☆ (5/10). Wrong!. (1/6).
1
3
29
My whole life, I'd never played Dungeons and Dragons, but I always figured it would be a guilty ultra-nerdy pleasure. I finally tried it and. all the numbers and dice are basically fake? It's really a bunch of people making up a story together?.
8
0
37
My tragically overlooked system for converting temperatures by transposing digits.
0
0
33
Dying for more aspartame content? Anyone?. I tried to lay out where the two recent WHO groups actually disagreed (not that much!) and why.
2
2
28
My more-hardcore theanine self-experiment . Here is a post! With words! Which is more nourishing than internet drama with plagarists!
1
0
30
How to Title your Blog Post or Whatever . 17 simple rules for naming stuff well. 🤯🧵
1
2
31
@cremieuxrecueil You're clearly very practiced at deflection. I'm not just accusing you of stealing my ideas. You also stole my words. That's plagiarism. If you like being accused of that—don't do plagiarism.
0
0
30
A study is a great idea. But let me pre-register my doubt that it's the walking. Arithmetic is good:.- Walking for an hour burns ~100 calories. - 1 lb of fat is 3500 calories. - Even if you walk an extra 4 hours a day that's only 3500/(4*7*100) = 0.8 lb of fat per week.
ok at some point we need to put together a study to systematically collect all the "I lost weight on a vacation" data points.
14
1
28
You’re in love. The two of you want to share the rest of your lives. So, being good game theorists, you have a romantic dinner and plan how to align your interests for mutually beneficial optimal strategic behavior. (1/8).
3
1
28
This seems like a good time to self-promote my rant on why (conditional) prediction markets don't generally give you causal effects.
4
0
29
@tracewoodgrains @cremieuxrecueil Shall we agree that @tracewoodgrains will determine the truth of this matter?. (And @tracewoodgrains can you confirm that you don't know either of us and have stake in this game?).
1
0
25
Someone, anyone: Choose left or right with a coinflip and apply retinol cream to that side of your face only for six months.
is anyone crusading on the lack of an evidence-base in skincare yet? . seems like low-hanging fruit to run some year-long RCTs and then start a website that sells only proven products with the data. unless the entire industry is a scam and you can use eucerin for everything.
8
1
24
0
0
25
The first RCT for GLP-1 drugs and alcoholism isn't what we hoped
5
4
24
@tracewoodgrains @cremieuxrecueil It seems we may not get any reply. @tracewoodgrains, I respect the writing you've done and would welcome your verdict. I'd be a fool to commit to abide by it unless @cremieuxrecueil does the same. But if you say this isn't plagiarism, I'll at least complain more quietly.
1
0
23
TWITTER PLEASE CALM DOWN. The claim is that blood PRODUCTS make up 2% of US exports by value. Much of the value of those products is contributed by things other than blood.
1
0
24
A modest proposal:. If you're a Democrat upset about being in this situation, how sending some appreciation to Dean Phillips for being LITERALLY THE ONLY PERSON willing to stick their neck out when it really mattered?.
1
1
20
Note sure if this is obvious or contrarian or both but here's an argument that lists are an objectively good way to structure writing:
3
0
21
Fun fact of the day: Private equity groups have acquired almost all of the firms that used to do independent IRB review for clinical research. Just two now account for 92% of applications to the FDA.
1
3
23
@tracewoodgrains @cremieuxrecueil Just wanted to make sure you've seen these comparisons. (There are many more I could provide.).
There's an amusing irony in the fact that if I link to my post the algorithm will suppress it. But I mean, compare these.
2
0
23
Are we likely to hit any barriers in scaling LLMs to larger datasets or to larger models? I did some estimates for that:
3
5
20
@moultano I'd say they MIGHT be bad for you. They could be totally harmless, I just think the risk/reward is not good.
0
0
23
I got so many amazing mentoring applications that I decided to choose randomly. I've been trying to email everyone. But in 2025, it's no longer possible to email 300 people? Even in chunks of 20, I get blocked as spam. So. I'm sorry if you haven't heard from me!
2
0
22
Prima facie, it's odd that there's no supplement company with a business model of, "We charge three times as as much but also we exhaustively test EVERYTHING". I guess this doesn't actually work? Maybe not a large enough market to amortize the testing costs over?.
3
0
21
Time for some more UNDERRATED REASONS TO BE THANKFUL. 1. That Earth is hot—maybe half from radioactive decay and half from leftover heat from when the planet formed—and heat is atoms jiggling around and the faster they jiggle the more often electrons absorb some kinetic.
1
0
20
Has someone already written "aluminium and health: much more than you wanted to know"?. (All I can find is kooky conspiracy theories and high-handed rebuttals that don't really engage with the evidence.).
2
0
22