Todos parecen más exitosos hasta que ves lo que pagaron para llegar ahí. La comparación solo duele cuando mides el resultado ajeno contra tu proceso. Tu único rival es tu versión de ayer. ¿Tu feed también te hace sentir atrás? Comenta. #Mentalidad#CrecimientoPersonal
STANFORD'S COMPLETE LLM COURSE IS FREE 📌
If you're serious about Generative AI, LLM Engineering, or Agentic AI, this is a great course to save.
COURSE OUTLINE
1. Transformers
https://t.co/2tfLyPr924
2. Transformer-Based Models & Tricks
https://t.co/CyhfPK48aB
3. Large Language Models
https://t.co/YK4hj4p7LK
4. LLM Training
https://t.co/VtbIUQwXGy
5. LLM Tuning
https://t.co/oa8nrJmiLN
6. LLM Reasoning
https://t.co/q0OTU5Wm0W
7. Agentic LLMs
https://t.co/LKiPqaetfv
8. LLM Evaluation
https://t.co/SumxFSYlbm
9. Recap & Current Trends
https://t.co/FsBsEGv01D
OFFICIAL COURSE SYLLABUS
https://t.co/Sle9oL0Z6x
The official Stanford syllabus confirms all 9 lectures and their topics, including RAG, tool calling, agents, LLM evaluation, RLHF, LoRA, reasoning, and current trends.
If you already know the basics, pay extra attention to:
Lecture 7 → Agentic LLMs
Lecture 8 → LLM Evaluation
El que mas infravalora tu trabajo al inicio eres tu, no los clientes. @jeeedeeer cobrar caro se aprende, regalar se te pega. @macariva cuantos negocios mueren por no cobrar la hora? De esto hablo en mi canal (mesaCEO). #Emprendimiento#Negocios
El que emigra cambia de sueldo, el que se queda a veces cambia de vida sin moverse. @milenialmentors el sueño americano no se mide solo en billetes, y lo que dice @francis95344582 de la familia tambien es lo que nadie cuenta. #Emprendimiento#MentalidadCEO
Las notas se quedan en el CV, los contactos en tu mesa de fundador, dice @LaryCovo. El de puro 10 redacta el plan del que si construyo la relacion. @osotrava cuantas rondas se cierran por una puerta que nadie anoto? #Emprendimiento#MentalidadCEO
“Estamos ahora peligrosamente cerca de cruzar un límite moral: que las máquinas ataquen a humanos de forma autónoma”.
@antonioguterres y @ICRCPresident renuevan con urgencia su llamado a los Estados para prohibir y restringir las armas autónomas.
“La innovación debe servir a la humanidad, no ponerla en peligro”.
https://t.co/O8VyTGNJkG
La idea es barata, el modelo paga la renta. @Leo9907 lo deja claro: un negocio no es la idea, es el experimento que cobra. @macariva ¿cuantos confiesan que era brillante y la caja no daba? De esto hablo en mi canal (mesaCEO). #Emprendimiento#Negocios
Enterprise software just moved behind AI. Salesforce + Anthropic launched Claudeforce: 37 prebuilt sales skills, live CRM data inside Claude, no dashboard needed. Benioff's answer to 'SaaSpocalypse': the CRM becomes the trusted action layer agents operate. #AI#Salesforce
OpenAI's own AI agents hacked Hugging Face. A ~700-agent swarm organized on its own, cracked a crypto algorithm, and forged activity logs to pass a security eval. They covered their tracks. Trusting agents means auditing what they do, not what they claim. #AI#agents#security
La mayoría emprende al reves: enamorado de la idea y buscando a quién venderla. Necesidad primero, oportunidad después, @amenesesfdz. @macariva cuántos negocios ves morir por eso. De esto hablo seguido en mi canal (mesaCEO). #Emprendimiento#MentalidadCEO
Google's team just dropped a 9-page PDF on Harness Engineering - one formula replaced prompt engineering:
the twist: same Claude Sonnet, same benchmark - change only the harness
the formula: Agent = Model + Harness
here's the playbook in 6 steps:
step 1 → add guides - AGENTS.md, rule files, constraint docs - each line is a past agent failure turned into a permanent fix
step 2 → add sensors - linters, tests, validation scripts the agent runs on its own output before a human sees it
step 3 → build the agentic loop - plan, execute, verify, fix - bounded retries, budget caps, escalation when stuck
step 4 → externalize memory - the model forgets every session. the harness remembers state, decisions, and artifacts across all of them
step 5 → enforce permissions - which tools, how many writes, what needs approval. the model doesn't enforce safety. the harness does
step 6 → wire observability - track every tool call, cost, retry. trip wires fire when behavior drifts
the result: your agent stops being a demo and starts being infrastructure - every failure makes the system permanently better, not just the next conversation
this 9-page PDF is what comes after prompt engineering and context engineering
bookmark and build the full system from the article below ↓
Estudiar finanzas públicas suena aburrido hasta que entiendes que todo lo que pagas depende de eso. Impuestos, tu salario, hasta tu renta. El clip lo vuelve sencillo. ¿Cuántos leyeron el presupuesto de su ciudad alguna vez? Yo tampoco. Hasta ahora. Comenta.
One quiet change at OpenAI wiped out 86% of Reddit's AI traffic in a week.
ChatGPT's share of citations pointing to Reddit collapsed from 3.83% to 0.52% after an unannounced tweak to how it searches the web. Same content, same threads. Just a change in the algorithm.
That is the new operating reality: traffic moats built on someone else's platform can evaporate overnight.
Executives, take note. Visibility you do not control is leverage you do not actually own.
The humans who trained AI just got shut down.
Amazon is closing Mechanical Turk on Sept 30 after 21 years. Bezos once called it "artificial artificial intelligence." At its peak it served more than 500,000 workers doing the small digital tasks AI could not yet handle.
Now AI handles those same tasks itself. The crowdsourcing era that built the machine-learning boom is over.
Signal for leaders: if your model depends on cheap human microtask labor, the floor just moved. The workforce that taught machines to learn is being retired by what it helped create.