Full disclosure, I worked briefly with Ello / Catalin some years back.
They've done a lot of work on at least two dimensions: (1) handling the nonstandard sounds and habits of speech of very early readers, who might be as young as four, and (2) connecting this to a specialized teaching system based on the science of reading, e.g. decodable readers.
The explanation I've heard is that the national curriculum switched away from phonics and to a teaching method called "balanced literacy", and this went terribly [1]. IIRC it involved teaching kids to recognize words by their overall visual shape. I believe this is what mature readers do, but the problem is that for this to work, one first needs to have bootstrapped a robust vocabulary. And that happens via being able to sound out words, i.e. phonics.
(I'm not sure whether Cato is generally reliable, but FYI there are lots of other writeups online on the same topic. It was the first non-paywalled and reasonably complete one I found)
Full disclosure: I worked on a small project with Ello / Catalin a few years ago.
At the time of writing, the sentiment in this post is that this is a terrible idea, and that kids need human tutors. The latter is 100% true. But also, you might want to know some facts about the state of children's literacy in the US (Ello is a math and reading tutor):
1. We're in crisis. As of 2025, 40% of fourth graders are reading below basic levels [1].
2. There's a massive teacher shortage. 2025 US state data shows ~400k teacher positions either unfilled or underqualified [2] – over 10% of the workforce.
3. Bloom's 2-sigma shows that 1-1 tutoring delivers outcomes at the 90th percentile of classroom teaching. Early research is finding that AI can deliver some of this benefit [3].
4. This can't always be solved by parents: 54% of US adults have a literacy below a 6th-grade level, and 20% are below 5th-grade level [4].
At Ello, I heard stories of children figuring out they were behind at school, and when given the app, they holed themselves up in their room and used it to get themselves caught up. And then they could read! Can you imagine falling behind at this critical juncture, and being stuck illiterate while your friends grow past you? We're currently setting kids up for lives of shame and deprivation.
My take: this really is a life-changing technology. And we need this problem solved. Democracy doesn't function without an educated populace.
Hello! I'm one of the main authors of the paper. Thanks for engaging with our work so thoughtfully – that's a very clear and valid question.
We didn't get around to addressing this within the paper itself – 80 pages is a lot, and deadlines, etc. But I have unpublished experiments that show that in a reasonably broad setting I'm doing some work in, verbalized probabilities are restoring a distribution that looks almost identical to the base distribution. It is not possible to demonstrate this on frontier models, since their public models are already mode-collapsed, and they don't share the base model or logprobs anyway. But I've established this to my personal satisfaction on large local models which offer base / post-trained pairs.
To share some intuition on why one might believe this is occurring: there are a bunch of tasks implicit in the pre-training corpus that encourage the model to learn this capability. Consider sentences in news and research articles like: "Scientists discover that [doing something] increases [some outcome] on [some population] by X%". It seems quite natural that the model might learn a pathway by which it can translate its base probabilities into the equivalent numeric tokens in order to "beat" the task of reducing loss on the "X%" prediction. I can even almost visualize how this works mechanically in terms of what the upper layers of an MLP would do to learn this, i.e. translating from weights into specific token slots. And this is almost certainly more parameter-efficient than constructing an entire separate emulated reality for filling in X. Although I'm not ruling out that the latter might still be happening – perhaps some future interp research might be able to validate this!
I'm actually working on a paper that packs up some of the above findings in passing. But if helpful in the meantime, this is also building on related work by Tian et al. 2023, "Just Ask for Calibration" [1] and Meister et al. 2024, "Benchmarking Distributional Alignment of LLMs" [2], that give some extra confidence here. Their findings indicate that whether or not they are rooted in the model's base probabilities, they seem to be useful for the purposes that people care about. (Oh, and you can probably set up an experiment to verify this independently with vLLM in a few Claude Code requests!)
Hope that was helpful – feel free to ping with follow-ups! (Although replies might be a little delayed, I happened to see this at a good time; having quite a crunchy week)
This is very funny to me. It took about a decade for them to receive a scientific name – because people were too busy eating them the whole time! The "note on the Bathynomus fishery" really makes the circumstances of this "discovery" quite clear.
Sadly, within the taxonomy itself the authors restrain themselves from sharing their findings on the most delicious parts and preparations of the animal. Darwin would have been disappointed [1], but at least as a species we've gotten our time down from 300 years [2].
I'm quite curious, what are people doing with AI-powered browser automation at the moment?
This is such a new capability that I'm having a hard time getting a sense of interesting use cases. I'm quite sure this is more than just a shim for web services which don't expose APIs. But I also wonder whether LLMs are good enough to be trusted with more open-ended tasks at this stage.
If the claim that the blogosphere is dying is true, does that imply the public intellectual commons is dying too? I suspect that while the cosyweb is more pleasant for most, this retreat might hinder vital testing and cross-pollination of ideas, and make it much harder for people to polarize into being intellectually active. For example, I've never been an active participant on ribbonfarm, but Rao's writing has made me a little smarter and inclined towards certain vectors of thought. And you can see ripples of his work in later writing by others.
What a shame it would be for this culture to be lost; while there's a lot of dross in the blogosphere, I don't know if the brightest jewels will still be possible in a future system of local, private, transient clusters of thought.
Thanks, I'm certainly going to try that. I was more asking if anyone has experience getting tests done properly in light of their low accuracy. From what I understand, an antigen test is still a stool test, meaning they are only 50% accurate. As a commenter on this post shared, managing the health system is challenging in this area. I just did a bit of googling, and found a couple of leads here:
> CDC recommends collecting three stool samples from patients over several days for accurate test results. Commercial testing products for diagnosing giardiasis are available in the United States. [1]
Perhaps running three tests is the standard of care, or if not one might advocate for this based on the CDC recommendation. And if dismissed, perhaps there are commercial products available at the consumer level.
Speaking as someone who went through this, running experiments with your diet is absolutely worth trying. It worked for me. There is actually a specific medical practice for this: look up "FODMAP". The idea is to temporarily cut out all likely suspects for a short period, see if that fixes things, and then gradually reintroduce them to identify the culprit. A gastroenterologist recommended this to me. It didn't help with my issues at the time as gluten is not covered by this cluster, but struck me as a very sensible approach.
In my experience the medical system is unusually useless and dismissive with digestive issues. I think this is probably related to how little it can do in this area. 10-15% of the US has IBS, and this is a disease of exclusion. That literally means that the medical system acknowledges a cluster of symptoms, but has no idea what is causing them or how to cure them. I can imagine that blaming patients is easier than the alternatives for some doctors.
Oh wow, I have the first half of this situation. I went through a period where my digestion was so bad that it was affecting my ability to function from day to day. I didn't get anything useful from my gastro; I even had a negative celiac antibody test. Eventually I started rigorously tracking everything I ate against my symptoms, and after a few months I was able to draw a strong correlation with gluten intake. From memory it was in the 0.7 range. The day I cut out gluten, a set of awful digestive symptoms completely left my life. They return any time I am glutened.
I was fortunate that over time I managed to return myself to full capacity, through reading a ton of research and running dozens of experiments like the above. But it was so damn hard. The symptoms reduced my ability to use my brain to fix myself. And if you're not a careful eater, it's not at all intuitive which foods contain gluten. This was also almost a decade ago while living in a developing country, so it wasn't even apparent that gluten might be a suspect.
I'm currently based in the US - does anyone know how one might get properly tested for chronic giardiasis, as a person who isn't themselves in microbiology? I almost certainly encountered poorly treated water in that period of my life.
Also - I can't help but suspect that a nontrivial percentage of the developing world is living below their full capacity due to something like this. Neglected tropical diseases are a horrendous category.
I'd love to hear more about your experience with this platform. Is there any public information about this platform or research? How did you guys manage the bad behaviors that come with anonymity, like trolling?
I'm currently feeling out a little research project around the use of AI in improving group decision-making, specifically with the hope of improving our political systems. There are so few IRL case studies, it'd be great to have more intuition for this!
That's very interesting - the Chinese "oi" has pretty much exactly the same usages, down to the little details! (I'm most familiar with the Cantonese 喂 wai2, though it's not my first language)
I wonder if that suggests this is one of those universal pre-linguistic words, like "huh" [1]. That feels especially believable given the utter simplicity of the mouth sound. It's almost just a yell, in the same way that "huh" is almost just an outbreath.
FYI OP is answering my question on what the historian's theory was! (They were downvoted)
This route makes sense to me; I do think the deal has been getting worse and worse lately, and periodic renegotiation is required to keep things running properly.
This is so insightful! What did they believe the main process was, that was causing things to break?
My guesses - this is about (a) inequality reaching a breaking point, where non-elites are starting to reject the social contract, making elite signaling dangerous, and (b) that elites at that moment don't have enough justification for their elevated status (mystification, misrecognition, etc)
> On flights within the U.S., airlines are responsible for lost-luggage reimbursement up to $2,500 per person; on international flights, airlines owe you a mere $9.07 per pound, with a ceiling of $640. (That rate was set by an international treaty in 1929.)
> Beyond that, airlines owe you nothing for your most valuable items. Most contracts of carriage specifically exempt from compensation things like antiques, art, books, documents, money, cameras, collectibles, electronics, or "fragile or perishable items." [1]
Also this is not a guaranteed payout, you may need to provide receipts for your things [2]. In general I would prefer to keep the contents of my bag vs receiving "up to" $2,500. We're mostly working people here. The time and money cost of travel disruptions, replacing stuff, navigating bureaucracy etc. is not low. And my stuff has emotional value.
I agree that this setup seems reasonable in a perfect world. But knowing how large, complex companies function, I feel it is unwise to create any kind of loop which rewards airlines as a function of luggage lost.
I agree in the abstract, but people and companies operate under bounded rationality. Customers can't easily price the expectation of lost luggage. Companies are made of subunits with independent budgets, each optimizing for incentives with short time horizons.
McDonald's is actually a good example, they did exactly this. IMO one reason their brand has been devalued is death by a thousand cost cuts. Each cut is imperceptible and seems like a win. But over several decades the net effect is a disaster. (IIRC Fast Food Nation documented some ideas about this process)
I think OP is right but hasn't fully unrolled what they mean. This seems like it creates a moral hazard [1], because it reduces the penalty to airlines for losing luggage.
If airlines were prepared to accept X in costs from lost luggage before, and now get paid Y, they can now accept about X+Y in lost luggage costs. Meaning more lost luggage! After the system dynamics resettle, from a certain angle this is basically stealing: customers pay the same ticket price, but are paying some percentage more in expected lost luggage.
And on top of that, this is just considering the financial cost - what is actually happening is that more travelers are being deprived of their personal belongings. These are worth more to people than insurance payouts! It's surprising, but when you zoom out this seems downright awful.