Alejandro Aguilera
183 posts

Alejandro Aguilera
@mrhappsmx
Me gusta la tecnología, hablar de ella y descubrir maneras de usarla.
Monterrey, Nuevo León Katılım Temmuz 2022
153 Takip Edilen38 Takipçiler

How can I make it so that Claude does not use 55% of a single session consumption in generating ONLY 2 documents?
I see this process going on:
> start skills to read
>process the whole context in each interaction
>generate code to produce the document
>execute code to produce the document
>convert output to PDF and image
>QA the output and validate through another processing of reading the image it has just created
>make corrections if something is not right
>repeat the cycle until it decides that the document is ready.
Why use this flow for everything?

English

¿Cómo puedo hacer para que Claude no utilice el 40% del uso en la generación de SOLO 2 documentos?
Porque creo que el problema está intrinsicamente relacionado al avance del modelo.
Para todo necesita las skills -> procesar todo el contexto en cada interacción -> generar código para producir el documento -> convertir a PDF y a imagen -> hacerle "QA" y validar que se vea bien a través de otro procesamiento de la imagen que acaba de convertir -> hacer correcciones si algo no está bien -> repetir el ciclo hasta que decida que el documento está listo.
Además para todo quiere generar artefactos, incluso para temas sencillos como crear un resumen.
¿Por qué usar este flujo para todo?

Español

@holafly_com hello I need help, I’ve had a bad experience, I don’t have internet all the time and specially now, I’m struggling a lot, I don’t have internet since yesterday. Please help, already tried basic troubleshooting like turning off and on my phone.
English

Will you address this? @willahmed I have 443 streak but regret subscribing another year and feel let down a lot by poor weight training analytics.

@levelsio@levelsio
Either @WHOOP starts tracking strength training properly or all of us lifters will just move to whatever wearable does @willahmed
English

I subscribed mostly for the “smart coaching” and “custom plan.” I wanted guidance on a realistic muscle-gain goal — how much weight, over what time, considering body fat and lean mass. But setup is just two questions:
Lose / maintain / gain?
Target weight?
That’s it. No body fat %, no timeline, no input on muscle vs. fat. It assumes you already know the plan. If you do, why pay long-term? Log for a few weeks, learn your numbers, and you’re done. Am I missing something?

English

I think we’re on a software revolution where digital transformation is key. Users are not finding Apple
Intelligence useful, but they’re not using anything as a substitute. Maybe Twitter will have us all believe that AI is everywhere, but I think 99% of people will still need to develop the skill of implementing AI into their lives and getting value from it. That is the challenge.
Now, granted, Apple needs to figure out a way to achieve this, but I think it will take at least another 2 years of refinement.
English

Perhaps that’s because Apple Intelligence does very little to nothing in value so far.
9to5Mac@9to5mac
Most iPhone owners see little to no value in Apple Intelligence so far 9to5mac.com/2024/12/16/mos… by @benlovejoy
English

@SnazzyLabs He is transitioning into his reptilian arc/form like zuck. Then will redeem himself after 2 seasons.
English

QUÉ ES ESTO? JAJA 😳
El nuevo modelo de voz de ChatGPT puede hablar en tono regio norteño 😂😂😂
#monterrey #chatgpt
Español

@MacRumors Where is this roadmap from? Or is it just speculation?
English

Pues muy diferente de la carretera Torreón Saltillo no está
Elon Musk@elonmusk
Home, sweet home
Español













