robwwilliams

LoL. We are all dancing naked on the table and hoping our clothes and wallet or purse and some of our pride will be where we left them.

No not necessary in my experience. But it can get pricey not to give Opus 5.5 some constraints.

Agree with afro88. Opus 5.5 as competent as Fable 5.1 on complex adversarial review of material and equations planted with errors. I still use both for a bit of variety.

Completely agree. I’ve had a Nature paper and a Nature Neuroscience paper published this year and the reviews from Opus were much more thorough and complete than those of the human reviewers. I still benefitted from the human reviewers, but now I consider an LLM review essential.

Ideally, but often not true of large complex studies. Typical example is that only a middle author statistician understands the details of data analysis. Also commonly true in quantitative genetics.

Actually I am in a very good position to tell. And who said “write for you”? How about “write with you as a collaborator?”.

Writing is a form of thinking and writing with an LLM makes the argumentation, discussion, and progress toward goals faster and often better in my situation.

I am stating an opinion based on 2.5 years of intense interactions and writing with Anthropic models on a difficult topic related to asynchronous computational systems.

Background: I am a neuroscientist and behavioral biologist (45 years of experience) who understands the biology of behaviors fairly well but I need to translate this background into hardware architectures that get beyond current asynchronous system (e.g., this recent neuromorphic paper in Nature Communication by Li et al., 2025, A deterministic neuromorphic architecture, doi.org/10.1038.s41467-025-65268-z).

Working AND writing with Opus and now Fable has been a blessing of sorts.

Too simple, too dogmatic at the start and at the end (“don’t use LLMs” and “then they are poison”). The middle section is all good.

Writing with a LLM is a boon if you are trying to write while mastering a new field. Imagine a neuroscientist entering robotics or an electrical engineer trying to enter the field of neuromorphic systems. In these cases a human will need help from an LLM to think and write more deeply AND broadly. I sure do.

You could object that this form of human-LLM collaborative writing is just for one reader; the writer. True at the start but not true toward the end.

I do agree that even the frontier models do not touch a really strong human writer. Way too much gimmickry. They do not yet match individuals such as Leslie Lamport, Terry Winograd, John MacArthur, Richard Rorty, Humberto Maturana, William James, Rodney Brooks. They don’t have the depth of understanding to simplify. But give them a few more years and excursions in the real world and a concept if “now” and the situation may improve.

I often ask Fable and Opus to explain difficult topics to me as if I were the canonical dog Rover. It works.

Yes, and genetic variant generally do not act in isolation. We currently focus on the small additive effects of variants because we can with small sample sizes—and 1 million humans is marginal using a GWAS cohort to dive into epistasis. But these interaction effects among variants are critical. Now almost completely deprecated.

van Creveld is correct spelling.

Poor commentary with no documentation or citation. At least point to the source of data for 200 plus countries. Presumably from IHME at the University of Washington.

This paper is worth reading. The caricature is not.

Global, regional, and national trends in the morbidity gap and contributing diseases, injuries, and risk factors, 1990–2023: a systematic analysis for the Global Burden of Disease Study 2023,” The Lancet Public Health, 2026, DOI: 10.1016/S2468-2667(26)00098-8

Also lovely new longitudinal data on centenarian heath across lifespan for a very large Swedish population that demonstrates persistent morbidity advantages compared to very well matched non-centenarians. Fun papers too: DOI: 10.1016/j.eclinm.2025.103396

DOI: 10.1093/geroni/igaf050

DOU: 10.1371/journal.pbio.3003784

And these conditions account are major contributors to the change: depression, anxiety, hearing loss, and low back pain. All are highly subject to diagnostic drift.

I have started beating the hell out of Claude with stylometric analyses of writers I admire; usually technical writers like Terry Winograd, Leslie Lamport, Rodney Brooks. Then the “no or minimal rhetorical flourishes” rule; no British em-dashes, and minimize the negative phrase thesis-antithesis fun”. It helps.

I think you are right that in a year or two LLMs will be able to do a good impression of many technical styles. But not Nabokov, Kundera, or Kafka for subtlety.

If one manages to channel Edsger W. Dijkstra I will be impressed and rank it high on my leaderboard.

Yep. Such a classic Claude flavored product. Good ideas wrapped in cotton candy.

I often read as much as 1000 words thinking to myself: “this is smooth”, but then think to myself “Is this my friend Opus 4.8 or now Opus 5?.

In this case the rhetorical neatness is unmistakable especially in the beginnings and endings of paragraphs: “Here is the constructive turn, honestly ranked, with no silver bullets on offer.”

Yes: and that is actually a smoking gun.

Easy enough to change. I have a Stylometry Skill fit to my preferred style—a mix of me and Terry Winograd. Give Opus 10 of your best paper thst you wrote and tell it to build a model of your style. hHard to distinguish except I make way more typos.

You patent and protect (as best you can) the missing ingredients needed to get to AGI. Only half joking and also scary to contemplate. Qualcomm’s CDMA patent is on example. ARM and Texas Instruments two other examples.

There is no one-size-fits-all, even for programming.

For my quirky workflows I need 100K even to prep Opus 4.8 or higher to begin good work. Not programming, but a complex synthetic architecture for implementing more attentional autonomy for machines.

Not how I view it as a 74 yo. Patients get what their doctors want as part of a culture of minimal preventive care (it does not pay) and massive medical care (procedures pay handsomely).

Try to get a unilateral diagnostic mammogram. Sorry in the system I am in there is no code for a unilateral diagnostic, only bilateral, even if only one side actually requires diagnosis.

Why? Because 2X the charge and income fir little extra care cost. And who would ever complain about such “excellent” care? Recent experience.

Funny: This title is a classic statement of Martin Heidegger’s. Go programmers!

Share your concern.

One point is that this is not “mind reading”. This is “motor command reading”. This is really no different than placing recording electrodes near motor neurons that direct vocalization. Or in terminal nerve fibers that move finger and wrist muscles.

For Meta this is a PR stunt of the type that I saw staged at a Human Brain Project demo in Budapest in 2012 in with an EEG-instrumented operator tried to move a robot around the stage with HBP leaders trying to narrate the “wonders of it all”. Sadly it was one of those classic demo failures in front of 200 neuroscientists and EU science bureaucrats. Didn’t stop the EU from funding the application the next year. About 1 billion euros not well spent.

In academic research I have seen this same trend, particularly intensely in IT and security. Lock it down, lock it up, and slow down research. A hierarchy of admins and techs taught to cover their asses.

I initial thought this was a great tongue-in-cheek comment. I still think it is a joke.

Not how I see it. And I have seen Europe close up for 60 years. It is wealthier and happier than it was even 15 years ago. Visit Spain, Portugal, Greece, and of course the other larger northern states; no comparison. Immigration has served them well in wealth and health. Compare Europe to the hollowed out core of the USA and you will shudder. Visit rural America.

The European “problem” is they live to have a life whereas in the USA we seem to live to work.

Yes, sure, but not efficiently. Even Pops will not want to run four hair dyer GPUs 24-7 in the garage.

I think you are right when you factor in the much more efficient newer high end GPU. That is what may make the current GPu investments obsolete in 2-4 years.

Yann is on the mark. Almost amusing to see the EU along with its many former “subjects” realize they are at great risk of joint Chinese-American hegemony in AI. We should all be more terrified of a few nation states defining the agendas and policies of AI use than current Ai variants that a inherently without purpose or autonomy.

Great analogy to the fear of the printing press being really bad news in that it enabled the rabble to get aroused.

Completely agree. This is what Hofstadter means by a strange loop. Our current LLMs have no attentional autonomy by design. The recursion is superficial and without its own Now. Adding attentional autonomy is The frightening alignment issue.

You nailed it. Asking the question is asking to define from the outside what is an inner recursive process. The question is a simple confusion of domains. This is Humbert Maturana’s main point in Autopoiesis and Cognition (1980, now reissued). Recommend the whole book, as does Terry Winograd. The most intense part is the appendix specifically about the nervous system. Nagel and others knew no neuroscience and are clueless about recursion.

Great, and won’t we all be just as surprised when human self-attentional control turns out to be just as simple or just as complex! Our minds as a strange fabric built of threads of recursions without the benefit of any explicit clock.

Diderot’s Jacque, The Fatalist is a great example. Perhaps even moreso of endless recursion to the end.

Living downwind of Colossus I and Colossus II in Memphis has orders of magnitude more weight than even a convention large data center. On par with a large cargo airport like MEM (FedEx hub).