Calambre @calambreDon - Twitter Profile

We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks: Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better: Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better: Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better: Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work! In summary: - As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding. - Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.

karpathy's tweet photo. We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:

Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better:

Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better:

Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better:

Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work!

In summary:
- As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding.
- Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.

1K

51K

6K

74K

6M

Calambre @calambreDon

3 days ago

@egraciani estais cooking en españa 🔥

1

2

0

109

Calambre @calambreDon

7 days ago

vamos hacia una mediocrizacion extrema en todos los ambitos

Calambre @calambreDon

7 days ago

el siguiente paso de la ai slopificacion vendra cuando todo el mundo tenga acceso a operaciones esteticas baratas. Todo el mundo sera guapo etc y completamente uniforme

0

26

0

16

Calambre @calambreDon

7 days ago

el siguiente paso de la ai slopificacion vendra cuando todo el mundo tenga acceso a operaciones esteticas baratas. Todo el mundo sera guapo etc y completamente uniforme

0

26

Calambre @calambreDon

7 days ago

he hecho uno y mi madre me lo ha confirmado. slop confirmado (a mi me sigue pareciendo que estan muy guapos pero nos vamos a hartar en 3 dias)

Guli Moreno

@GuliMoreno

7 days ago

Soy al único al que ya todos los videos de Opus 5.5 le parecen tremendo AI slop?

29

93

0

6

10K

0

27

calambreDon retweeted

javi santana

@javisantana

7 days ago

"Una ley puede mejorar los incentivos. No puede fabricar ambición, referentes o confianza." En particular en España no te enseñan que si sabes hacer algo puedes decidir no trabajar para nadie. Yo tardé años en darme cuenta.

4

114

8

21

11K

calambreDon retweeted

David Pardo @dei_biz

8 days ago

@barckcode Más allá de esa dualidad que comentáis de programar <> crear, creo que también tiene que ver con la neurdivergencia y la capacidad de concentración. He visto que los que tenemos facilidad para distraernos e hiperfoco, gozamos como gorrinos. Los más estructurados sufren más.

6

39

3

1

6K

Calambre @calambreDon

8 days ago

@iagolast yo no pude tampoco 🥲

0

1

0

607

Calambre @calambreDon

8 days ago

"Asymmetric opportunities: Invest in startups Start a company Create a book, podcast, video Create a (software) product Go on many first dates Go to a cocktail party Read a Lindy book Move to a big city Buy Bitcoin Tweet." + apuntate a todo lo que Guli saca

Guli Moreno

@GuliMoreno

9 days ago

Aprovechando el tirón de nuevos seguidores por @hackspain, aprovecho para compartir el Manifiesto de @exp8fellowship. Para que se entienda mejor por qué existe Exponential. https://t.co/qaO3m1ZNmA

29

384

90

293

91K

1

7

0

7

2K