Newsletter
Bias vs Noise pt. II: Stress and other things
March 28, 2025
Hello,
Here’s everything since my last little missive to you:
Excerpt: Bias is just you using your expectations and assumptions to ignore the noise, and see the picture more clearly. The trade-off is that, sometimes, the noise is useful or your expectations are off. The human stress response is perhaps the most fundamental example of this in behaviour, and a very valuable tool.
Main idea: Stress promotes bias—stereotypical thinking and behaving. Less stress promotes cognitive flexibility—an openness to new ways of thinking and behaving. Neither is better than the other. It’s about the situation you deploy them in.
Interesting article on how AI thinks, by Anthropic. Some highlights:
Knowing how models like Claude think would allow us to have a better understanding of their abilities, as well as help us ensure that they’re doing what we intend them to. For example:
- Claude can speak dozens of languages. What language, if any, is it using “in its head”?
- Claude writes text one word at a time. Is it only focusing on predicting the next word or does it ever plan ahead?
- Claude can write out its reasoning step-by-step. Does this explanation represent the actual steps it took to get to an answer, or is it sometimes fabricating a plausible argument for a foregone conclusion?
And:
solid evidence that:
- Claude sometimes thinks in a conceptual space that is shared between languages, suggesting it has a kind of universal “language of thought.” We show this by translating simple sentences into multiple languages and tracing the overlap in how Claude processes them.
- Claude will plan what it will say many words ahead, and write to get to that destination. We show this in the realm of poetry, where it thinks of possible rhyming words in advance and writes the next line to get there. This is powerful evidence that even though models are trained to output one word at a time, they may think on much longer horizons to do so.
- Claude, on occasion, will give a plausible-sounding argument designed to agree with the user rather than to follow logical steps. We show this by asking it for help on a hard math problem while giving it an incorrect hint. We are able to “catch it in the act” as it makes up its fake reasoning, providing a proof of concept that our tools can be useful for flagging concerning mechanisms in models.
And:
In a study of hallucinations, we found the counter-intuitive result that Claude’s default behavior is to decline to speculate when asked a question, and it only answers questions when something inhibits this default reluctance.
Lots interesting. But I still think I disagree with calling this ‘thinking’. More here and here, but I still see no reason to believe this isn’t more like walking for an AI. On that account, these ‘thought processes’ would be more like adjusting to terrain, or something.
–
Main story is about a man who walks again, but others experience smaller improvements. The ‘reprogrammed’ cells:
are created by reverting adult cells to an embryonic-like state, from which they can be coaxed to develop into other cell types.
Feel-good.
–
Stolen from a tweet. Not groundbreaking, but interesting thoughts on use-cases for the new OpenAI image model:
(1) This changes filters. Instagram filters required custom code; now all you need are a few keywords like “Studio Ghibli” or Dr. Seuss or South Park.
(2) This changes online ads. Much of the workflow of ad unit generation can now be automated, as per QT below.
(3) This changes memes. The baseline quality of memes should rise, because a critical threshold of reducing prompting effort to get good results has been reached.
(4) This may change books. I’d like to see someone take a public domain book from Project Gutenberg, feed it page by page into Claude, and have it turn it into comic book panels with the new ChatGPT. Old books may become more accessible this way.
(5) This changes slides. We’re now close to the point where you can generate a few reasonable AI images for any slide deck. With the right integration, there should be less bullet-point only presentations.
(6) This changes websites. You can now generate placeholder images in a site-specific style for any
tag, as a kind of visual Loren Ipsum.
(7) This may change movies. We could see shot-for-shot remakes of old movies in new visual styles, with dubbing just for the artistry of it. Though these might be more interesting as clips than as full movies.
(8) This may change social networking. Once this tech is open source and/or cheap enough to widely integrate, every upload image button will have a generate image alongside it.
(9) This should change image search. A generate option will likewise pop up alongside available images.
(10) In general, visual styles have suddenly become extremely easy to copy, even easier than frontend code. Distinction will have to come in other ways.
–
Article on the health cost of the current economic malaise. US-centric, so you’d have to modify by satisfaction when doing the hypothetical in your own country. Feels like the UK might be doing worse, to me. Horrible time here:
a significant increase in mortality among middle aged white men and women — an increase concentrated amongst lower income, working class Americans.[ii] Case and Deaton trace the proximate causes of death driving this increase to suicide, drug and alcohol poisoning, and chronic liver diseases and cirrhosis. Measures of self-assessed health status they examined in surveys over 2011-2013 compared to 1997-1999 also show increased reports of pain and psychological distress.
And, the comparison case:
These results bear a striking resemblance to another demographic crisis: Though we are used to thinking of the Cold War as an economic and political contest without casualties the fall of the Berlin Wall showed us that when economic systems and expectations collapse, people die just a surely as they do in a shooting war. In the early 1990s, in the aftermath of the collapse of the Soviet Union, life expectancy in the former Soviet Union and Eastern Europe fell dramatically.
Worth being rich.
–
Why We Ended Up Homeschooling. Anecdotal, but I liked this line:
I don’t think most people understand the true opportunity cost of schooling because they have never seen what is possible.
Worth exploring what’s possible.
–
SSRN article on the extent to which we underestimate skill decay:
we investigate the accuracy of beliefs about skill decay. Participants consistently underestimated their own skill decay by 28% to 59% across tasks. Even after directly experiencing skill decay, participants continued to underpredict its extent. We identify two mechanisms driving this underestimation: First, participants were more accurate in predicting others’ skill decline than their own, suggesting ego-based motivations are at play. Second, both subgroup heterogeneity and variable importance analyses reveal an underappreciation of the adverse impact of age on skill decay. Together, these findings suggest systematic misjudgments of skill retention, with implications for human capital investment decisions.
–
Are the kids better at all? I complained a lot with fellow teachers that students seem to have taken a real hit in terms of engagement in the last couple of years. Easy to complain about the kids being worse, but this tweeter is collecting examples of where they’re better. Music, sport, and technical specialisation seem to be the common theme. Probably primed, since I used my own poorer engagement anecdote as the lede, but I wonder if this is a trend away from synthesis.
–
I hope you found something interesting.
You can find links to all my previous missives here.
Warm regards,