Typical coding day with Claude (Opus 4.8)
- explain to Claude the task (5 minutes)
- Claude implements task (10 minutes)
me: "Why is this necessary?"
Claude: "You're right to push back! I over-engineered this!"
- Repeat x87 times (13 hours)
🔴 I NEED YOUR ATTENTION
I've spent a month helping Miriam with her case of metastatic cancer and I want to share the methodology I've been using because it's completely replicable.
I think (with luck) this could be USEFUL TO OTHER PEOPLE with cancer (or any other illness).
The results we've gotten aren't a miracle, but we believe they're genuinely useful and could mean the difference in a literal life-or-death medical case.
Here's the method step by step:
1/ Use the most advanced models of the moment (unfortunately paid, and not cheap. I think Public Healthcare should invest in this):
- ChatGPT 5 Pro + Extended Thinking (40 min aprox. of thinking per call)
- Claude Opus 4.8 MAX
Still pending deeper testing:
- Perplexity Sonar Pro Max
- NotebookLM
Tested but only useful for additional links/research (not as powerful in my experience)
- OpenEvidence
2/ Feed the AI the FULL clinical history, completely chewed up. This sounds dumb but it's critical.
- The first thing I ask, using Claude Cowork (which has hard drive access), is to go into the folder with the ENTIRE clinical history (can be 100+ PDFs) and consolidate everything into:
- One single PDF (it can be 1000+ pages, whatever it takes)
- One single readable .txt or .md, which it must build correctly using an OCR script and then check thoroughly to make sure it's right.
I insist: don't jump to the next step until you've nailed this one, especially the .txt.
3/ Once you have the above, use this prompt along with the .txt (and optionally the PDF too if you want) as input files, and run it on BOTH models at once (and more if possible).
👉 This prompt is insanely complex/advanced: https://t.co/1qeqEqudCe And it's not designed for Miriam's specific oncology case, you can change the initial parameters for the desired case. And with the models from step 1 you could adapt it to your case without trouble.
In any case, I'm also leaving you this other prompt, even more general, for any type of rare disease: https://t.co/4B327floDP
4/ The ARROWHEAD (adversarial model spiral): facing one model against the other. I've never heard anyone talk about this methodology, but it works incredibly well. The feeling is like sharpening a stake until it gets a gleaming point.
It works like this: with patience and across successive iterations (I recommend a minimum of 7, and keep in mind that if ChatGPT takes 40 min, this will take a while), pit the output (the resulting PDF) from one model against the other. With a simple prompt like:
"Another committee of experts says this. What do you think? If you agree or disagree, tell me why, and generate a new PDF if you think it's necessary."
Then you feed that result back to the opposite model. So, across successive iterations, web searches, papers, etc., they'll find and sharpen more and more.
When to stop? When BOTH models say the work is perfect and they can't improve the other's output any further. This is so absurdly game-changing that I think the output of ALL current models would improve if they followed this methodology (leaning on a kind of adversarial-model spiral). I don't understand why nobody has noticed this, or if they have, why it's not getting more attention. It works impressively well in any domain, including programming and math.
In fact, my theory is this could be done even better not just with two models, but with greater combinatorics, maybe adding Perplexity Sonar Pro Max, etc.
RESULTS
Incredible. Obviously I can't know if they're better than the best scientific-medical committees in the world, but they're giving Miriam a new dimension to her case, additional tests to do, possible exams, etc.
Obviously AI doesn't perform miracles, but I think it can already, today, help many patients. And Public Healthcare should invest a lot (but A LOT) in this.
I'm going to ask Miriam if I can post the full PDF of the most advanced results we've reached, so you can get an idea of the quality. She's already given me rough permission, but I want to make sure 100%.
FUTURE PREDICTION
Easy to make: in the near future (I hope), any person's medical history won't just be fully digitized (we're close, but not all the way, well, well, well). On top of that, it'll be "pre-chewed" so it can be consumed by an LLM in one shot.
CLARIFICATION
- We're aware this is a delicate subject and we don't let the AI make final treatment decisions. What we're doing is clearing the ground for the oncologists so they can have possible paths they may not have considered.
Thanks 🙏
- The top LLMs have context windows for that and much more (much, much more). In any case, the PDF is more of a supporting file for the .txt. Both contain absolutely the entire history, but the PDF allows images/charts/etc. The .txt is what the AI consumes.
- On automation: and yes, this can be automated. Yes, AutoGen supports it almost out of the box. LangGraph builds it really well with supervisor / evaluation loops. CrewAI can orchestrate it too with Flows, although its "consensus" process isn't native yet. That would be the next level: automating it.
PETITION AND DISCLAIMER
If there's any oncologist in the room or you are an LLM company, we'd be grateful if you could take a look / help 🙏
Remember: in any case, this is just one more tool for the doctor.
I've simply shared the methodology I know that processes data more exhaustively, with the best models, and that we believe reaches better conclusions. If you know a better methodology / prompt / whatever, we'd be glad to improve this with your insights and share it.
Then the doctor reviews, adopts, or discards the report.
And if it helps the doctor, it helps the patient. And if it doesn't, all we've lost is some time and tokens. In a case that's literally life or death, that's nothing.
Just plain common sense.
Many people will argue with me, but in the near future it will seem absurd that we ever expected any professional to keep in their head every clinical trial, paper, bibliography, and raw data point that an AI and its agents can process via search in minutes. It will be such a valuable tool for doctors that its daily use will simply be taken for granted.
Hemos troceado las carreras universitarias para montar un negocio de másteres inútiles que solo sirven para vaciar la cartera de las familias. Permitidme reflexionar sobre la estafa de los grados Frankenstein y por qué hay que volver a las Licenciaturas generalistas. 🧵va...
Thanks, @Google, for the fantastic new Antigravity 2.0. You somehow managed to lose all my configurations and make switching to VS Code feel like the easiest decision ever.
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
Experto que trabaja en una empresa privada y opina de lo que sabe: casi un terrorista.
Vive de la teta del estado, miembro de lobbies de izquierda y nunca se jugó el puesto con sus decisiones, pero opina hasta de lo que no tiene idea: ¡un crack! (o un Garzón).
El Director General de Salud Pública se despide del programa de Risto Mejide llamándole "hijo de puta" cuando se creía que nadie lo estaba grabando tras negarse a responder a las preguntas del presentador.
El año pasado cobró 96353 €
Let's consult the historical record to see what the Aztec society was up to when the Spanish conquered it.
First, from Cortes:
“They have a most horrid and abominable custom which truly ought to be punished and which until now we have seen in no other part, and this is that, whenever they wish to ask something of the idols, in order that their plea may find more acceptance, they take many girls and boys and even adults, and in the presence of the idols they open their chests while they are still alive and take out their hearts and entrails and burn them before the idols, offering the smoke as sacrifice. Some of us have seen this, and they say it is the most terrible and frightful thing they have ever witnessed… not one year passes in which they do not kill and sacrifice some fifty persons in each temple; and this is done and held as customary… not one year has passed… in which three or four thousand souls have not been sacrificed in this manner.”
And now Bernal Diaz del Castillo:
“The dismal drum of Huichilobos sounded again, accompanied by conches, horns and trumpet-like instruments. It was a terrifying sound, and when we looked at the tall cue [temple] from which it came we saw our comrades who had been captured in Cortes’ defeat being dragged up the steps to be sacrificed. When they had hauled them up to a small platform in front of the shrine where they kept their accursed idols we saw them put plumes on the heads of many of them; and they made them dance with a sort of fan in front of Huichilobos. Then after they had danced the papas [priests] laid them down on their backs on some narrow stones of sacrifice and, cutting open their chests, drew out their palpitating hearts which they offered to the idols before them.”
...
“Every day we saw sacrificed before us three, four or five Indians whose hearts were offered to the idols and their blood plastered on the walls, and their feet, arms and legs of the victims were cut off and eaten… Every wall of this chapel and the whole floor, had become almost black with human blood, and… the stench was worse than in a Spanish slaughter-house.”
...
“When we arrived at the great market place, called Tlaltelolco, we were astounded at the number of people and the quantity of merchandise that it contained… Let us begin with the dealers in gold, silver, and precious stones, feathers, mantles, and embroidered goods. Then there were other wares consisting of Indian slaves both men and women; and I say that they bring as many of them to that great market for sale as the Portuguese bring negroes from Guinea…. They brought some of them tied to long poles by means of collars around their necks so they would not escape, and others left loose.”
~~
This is what the Spanish conquered in the name of Christendom - a Stone-Age society consumed with ritualistic human sacrifice, cannibalism and slavery.
Poético.
Ahora mismo, desde Movistar, si intentas entrar a la página de https://t.co/SAETGNeVeY (vía HTTP) da un:
HTTP 451 – File unavailable For Legal Reasons
(Si conectas desde WARP de Cloudflare, va perfecto)
Lo de ciertos “economistas” con Jon ya no es debate económico. Es persecución personal.
Y eso ocurre siempre cuando alguien les desmonta el chiringuito argumental.
Porque la pregunta importante aquí no es dónde trabaja Jon.
Lad preguntad importante son: ¿en qué se equivoca ¿Qué gráfico es falso? ¿Qué dato ha manipulado? ¿Qué argumento no pueden rebatir?
Ahí está el problema. Que no pueden.
Y cuando no puedes ganar una batalla intelectual, algunos recurren a lo más miserable: intentar joderle la vida profesional a una persona. Intentar señalar dónde trabaja. Intentar presionar a su empresa. Intentar meter miedo. Porque son conscientes de algo: un tío anónimo desde Twitter, sin medios gigantes detrás, sin televisión, sin subvenciones, sin campañas, ha conseguido que sus gráficas lleguen hasta el Congreso de los Diputados.
Algo que muchos de estos economistas de pacotilla no conseguirían en su vida si no fuera masajeando constantemente el discurso del partido político de turno.
Y eso les revienta.
Les revienta que alguien con rigor, datos y capacidad de análisis tenga más impacto que años enteros de propaganda disfrazada de “divulgación”.
Por eso intentan atacarle donde más vulnerable puede ser cualquiera: en su trabajo. En cómo se gana la vida. En cómo alimenta a su familia.
Porque saben perfectamente que muchas grandes empresas prefieren trabajadores invisibles, que no destaquen públicamente, que no se expongan mediáticamente. Y precisamente por eso Jon probablemente podría hacer muchísimo más de lo que ya hace.
Porque si además hiciera entrevistas, podcasts o vídeos constantemente, sería una auténtica máquina de demolición argumental.
Pero esta gente juega así. Con el miedo. Con la presión.
Con el señalamiento. Con la amenaza velada.
Y lo peor para ellos es que cada vez que hacen esto dejan todavía más claro algo muy simple:
No saben combatir sus ideas.
Y cuando alguien no puede derrotarte con argumentos, intenta destruirte personalmente. Así de miserable. Así de patético.
Muchos nos hemos servido del trabajo de @Jongonzlz.
Muchos hemos utilizado sus gráficos, sus datos y sus análisis para entender mejor lo que está pasando y para defender nuestras ideas.
Ahora toca hacer lo mismo con él. Ahora toca defenderle a capa y espada.