Recommendations

I give the AI a list of my favorite books or music, explain what I like about them, and that’s it: I get a list of personalized recommendations. That’s how I’ve found musical oddities and a couple of great literary recommendations I probably never would have gotten around to reading.

Prompt for book recommendations

My favorite books are X, Y and Z, and I like them especially because of A and B.
Before recommending anything, ask me 3 or 4 questions to fine-tune my taste.
Then recommend 5 books I probably haven't read, explaining in one line why each one connects with what I like.

or, for music:

Prompt for music recommendations

I like the music of X, Y, ..., Z.
Before recommending anything, ask me a couple of questions to understand what draws me to them.
Then recommend lesser-known artists I might like, grouped by the style or reason that connects them to my taste.

Shopping assistant

When I have to buy something online, I tend to fall into a terrible decision paralysis. I over-research, build a spreadsheet with features and prices, and almost certainly spend way more time than necessary. Sound familiar?

Asking the AI to first help me define what I’m looking for, and only then to compare, saves me hours. Literally. The key shift was to stop asking it for “the best product” and start asking it to help me decide what “best” means for me.

Prompt for shopping

I want to buy X, available in Y. My budget is Z and I'll use it for [use case].
First, help me define the 4 or 5 features that really matter for my case, and ask me whatever you need.
Then compare the best options in a table with price, pros and cons, and close with your recommendation and why.

Finding your purpose

I asked Claude to review my LinkedIn, my website and other things of mine online, and to ask me questions to help me find my purpose (my why). It surprised me: it made a good summary of my activities and pushed me to think. That said, I’ll be honest: I had to iterate through several versions before I was happy with it. It didn’t come out right the first time, not even close.

Prompt for your purpose

Act as a leadership coach who's an expert in Simon Sinek's "Golden Circle" (what/how/why).
I want to identify my WHY: my fundamental purpose or belief.
Review my context [paste profile, links or background] and look for the common threads between my achievements, my hard decisions and what gives me deep satisfaction. Don't stay on the surface.
Ask me between 5 and 7 open questions about the "what for" of what I do.
When I answer, help me draft a clear, actionable purpose statement.
Start by asking me the first question.

Analyzing conversations

When I have to have a difficult conversation, I write down my points beforehand so I can review them calmly. I discovered that passing them to the AI helps me prepare better: it finds the weak spots in my argument, suggests analogies, and sometimes makes me see the other side in a way I wasn’t seeing on my own.

Prompt for conversations

I have to have a difficult conversation with [person] about [topic]. My goal is [goal].
These are my notes:
X
Y
Z

Read them from the other person's perspective and tell me: where my argument is weak, what could be misinterpreted, and how you'd rephrase it. If an example or analogy helps, use it.

I’m not sure you can even call this “productivity”. For me it’s more like using AI as a thinking partner: someone/something that forces me to ask more and better questions.

Attempt #1 - the proof of concept

The proof of concept helped me understand a lot of things.

I started by making the connectors for the rotating piece separately, but it was cumbersome to insert and the pieces fit very loosely. Also, the pieces were too wide (2 cm), which made it look very clunky.

Lessons learned: It’s easier to design a single model to print, and to make certain joints break to allow rotation.

Attempt #2 - styling

The top part is printed as a single piece, which has supports that break elegantly to allow free rotation. By reducing the width of the pieces, the trophy looks more stylized, but it also looks fragile. The joint between the base and the rotating plate looks very weak.

I didn’t like the surface finishes. The “rotated” surfaces on the base create a certain dissonance for me. What I learned: To design supports integrated into the model (instead of having them generated automatically).

Attempt #3 - final

The joint between the base and the rotating plate still seemed weak to me, so I added an extra piece that allowed a snap-fit joint to provide greater stability. I printed the trophies with black and “gold” filaments.

Friday

On the first day, walking through the convention center, I was surprised by the size of the event. The rooms are gigantic. Truly enormous. My hotel -expensive but sponsored by the PSF- had a direct connection to the event center, which was very convenient.

In the photo, you can see the main space on the 2nd floor, which had posters, booths from companies, startups, and communities, and the food service area. On the same floor was the main hall where the keynotes and lightning talks took place. On the 3rd floor were the rooms for the talks and the charlas. Yes, there is a difference. There were four rooms for talks (talks in English) and one room for charlas (talks in Spanish). Additionally, on the 3rd floor, there was a green room, a space for speakers to test their connections and practice. As a speaker, I had the right to go there whenever I wanted, which was convenient for getting coffee when it had run out elsewhere.

In the morning, I registered for the event and got my badge with my name, stickers, and a t-shirt. The t-shirt was way too small for me (L fitted << regular M), but fortunately, I was able to exchange it on Sunday. Even during breakfast, it was clear this wouldn’t be just any event: in the distance, I saw Guido van Rossum, the very creator of Python. Yes, he’s also human. Incredible. I resisted the urge to take a picture because I had already been warned not to paparazzi him.

In the morning, I met up with Denny and Cristián, whom I have talked to and collaborated with online through Python Chile for years but had never met in person. And it felt like we already knew each other and had always seen each other. I was at the Spanish Python booth, meeting other Python enthusiasts, giving out stickers, and receiving stickers. I also attended the PyCon inauguration and the first keynote.

I also stopped by the Streamlit booth and took the opportunity to greet Chanin, Tony, Cristine, Ted, and Krista, whom I had only known virtually. I was surprised by the warmth with which they greeted me with me. I would have liked to talk more with them and express my gratitude, but I think the mix of fatigue, lack of sleep, and a constant state of amazement limited my ability to express myself in English. It was only on Saturday, after my talk, that I managed to relax and truly enjoy it.

Chanin - Data Prof - interviewed me trying to get some coherent thoughts.

After the keynote, I split my time between wandering around the stands, collecting t-shirts and stickers, and attending some talks. There was a stand with puzzles to solve and t-shirts that I liked, but it was always very crowded, so I decided to try again the next day.

During the lightning talks, around 6 pm, I realized I was mentally and emotionally exhausted. I decided to go for a walk to relax and get to know the surroundings a bit, and to buy something for dinner. I realized I needed to rest to be able to give a good presentation the next day. I ended up having dinner at the hotel despite walking around looking for a better option (3 mini-burgers). Back in the room, I rehearsed and worked on the last changes to the presentation from 8 to 11 pm. My speaker notes had every word I needed to say and even the moment to click/next on the presentation. I collapsed into bed and slept deeply.

Saturday

The day started early, with a light breakfast due to my pre-presentation nerves. I went to Simon Willison’s keynote, which was incredible and I will recommend to everyone as soon as it’s available on YouTube. Simon explains how LLMs work in a didactic and fun way, highlighting all their limitations but also the amazing things they allow us to do (which weren’t possible before).

At noon, I gave my talk in English in a huge, packed room. To be honest, I was incredibly nervous. Just before going on stage, I realized that I didn’t have a single line of Python in my presentation! Fortunately, it was too late to change anything. I knew I had prepared good material and notes to support me.

The reception has been overwhelmingly positive. I think storytelling really connects with people because it is a very intuitive and powerful concept. We all want to improve the way we communicate our ideas. Hearing people say: “Thanks for pointing out library X, I love it” or “I wanted to keep learning, so I bought book Y” makes me very happy.

In the presentation, I included a survey, which over 30 people responded to. The feedback is overwhelmingly positive, and the few suggestions for improvement are:

• I need to smile more
• I need to modulate better and use more tones, not speak so flatly
• The ending was a bit abrupt
• The agenda (table of contents) can be used to generate more interest

After the talk, I took the opportunity to relax and attend some presentations. I also chatted with some people who approached me to comment on my presentation, exchange book recommendations, or simply introduce themselves.

Do you remember the stand with the t-shirts I wanted? In the afternoon, I stopped by and it was almost empty. There were three different challenges, and by solving each one, you would get a four-digit code that you could use to open a lock on a box. While they were explaining the rules, I half-jokingly asked if it was allowed to open the lock without solving the puzzle, and they looked at me like, “Can you really do that?” So I showed them that it’s actually super easy to hack one of those small locks (I had seen a tutorial a few years ago) and managed to open the lock in about a minute. They laughed a lot and gave me the t-shirt anyway (I never got to learn what the puzzles were about). I felt so “street smart” haha.

Like the previous day, I alternated between attending talks, visiting the stands, and networking. I met a lot of interesting people. At lunch, I met a Duolingo engineer who worked in Pittsburgh and a recent graduate who was just starting to learn Python. The sheer diversity of profiles and knowledge. And I love that there is no prejudice about your level of knowledge, because there is so much to learn that we are all newbies in some aspect or library of Python. We are all always learning something.

I went back to the hotel to drop off my computer and dressed up for the PyLadies auction. The “PyLadies auction” is very fun and noisy, but also elegant and intimate. A multi-course menu with very good food is served. Meanwhile, various items donated by people from the community are auctioned off. Python Chile had donated one of the keepsakes from PyCon Chile 2023 (the elevator with a crank), which ended up selling for 425 dollars - I only hoped it would go for more than 100 dollars. There were 44 items up for auction, including: a giant plush snake (3 or 4 meters), a suitcase with the logos of all the PyLadies chapters around the world, and a cuckoo clock with a Python theme. In total, more than 60,000 dollars were raised.

After the PyLadies event, people from various Spanish-speaking communities and I went out for beers and to celebrate Cristián’s birthday. I think I went back to the hotel around 2 am.

Sunday

I woke up early because I had committed to presenting a slide about PyDay Chile during the lightning talks at 8 am. I enjoyed starting by saying, “¿Que tal, festival?” which surely only forty-something Chileans will understand as a reference.

By Sunday, there were fewer people at the event, with many walking around with their suitcases. The company, startup, and community booths had already been removed, and the focus was on the job fair and posters. I found the work on “algorithmic embroidery” notable for its artistic and delicate nature. It’s incredible that such things can be done, linking the ethereal nature of a programming language with the concrete reality of fabric and thread.

In the morning, I attended some talks, and in the afternoon, I organized an Open Space called “Data Storytelling 2: Revenge of the Audience” that I had proposed. I wasn’t familiar with the Open Space format, and it’s super powerful. In addition to the scheduled talks, about 10 rooms of various sizes (60 to 100 people) are available to attendees, and depending on the mood and interest, “collaboration spaces” are proposed, which can take any form or intention.

Many more people attended the Data Storytelling Open Space than I had expected; I think there were about 15 of us. We talked about different topics:

• Favorite data visualization libraries.
• Opinions on pie charts and other graphs.
• Using renpy to make a presentation that is like a conversation/manga.
• Using the manim library to make mathematical animation graphics.
• Book recommendations: Creativity Inc, which talks about storytelling at Pixar.

One of the most interesting questions was, “What will happen in the future when generative AI can create a movie, comic, or book from a prompt?” This will mean that as a society we won’t have the same references, and that they might not be transferable.

We also took photos of the Spanish-speaking Python community.

Armando and I went to dinner at Primanti Bros, a sandwich shop in Pittsburgh that is supposed to be extraordinary but pales in comparison to the sandwiches from Chile, Peru, or Argentina.

Afterwards, we met up with the rest of the Pythonistas at a nearby brewery. There, we talked, toasted until glasses broke, and gave away the last stickers, plush toys, and t-shirts. One Pythonista was looking from afar with a face that said, “I want a sticker too,” so I invited him over. He took some stickers for his daughter, and we started chatting. He was American but spoke Japanese and was very close with the Python Japan group, whom he introduced me to, and he ended up buying me some beers. Two stickers for two beers, definitely the best deal I’ve ever made! Hahaha. I was surprised at how easy it was to meet people and discuss technical topics when I usually find it much harder to talk to strangers or network.

And so, the time came to pack my bags. I decorated my suitcase with NumFOCUS stickers and headed to the airport, thinking it was true, that I had accomplished it. My dream of being at a PyCon US and presenting a talk in English. That when you think big and work from and for a community, amazing things can be achieved. And that I have been very fortunate to be able to get here, and that also means a huge responsibility to support others in daring to come as well.

Traditional methods

One option is to apply the methods directly:

df_g = df.groupby("color").mean().reset_index()
print(df_g)

   color         x         y
0   azul -2.666667 -4.333333
1  negro  2.000000  2.000000
2   rojo -3.500000  3.500000

or applying it to a specific column:

df_g = df.groupby("color").y.mean().reset_index()
print(df_g)

   color         y
0   azul -4.333333
1  negro  2.000000
2   rojo  3.500000

It’s not advisable because: * You’ll have to manually rename the column names, which keep the name of the original column and not of the applied function. * You can’t apply the method to more than one function and column simultaneously * You can’t use any function, only some well-known ones.

Another method that can be used, but isn’t as convenient, is to use agg:

df_g = df.groupby("color").agg(
    {
        "x": [lambda _: _.mean()], # applied to y
        "y": [lambda _: _.nunique(), lambda _: _.mean()], # applied to z
    }
)
print(df_g)

              x          y           
       <lambda> <lambda_0> <lambda_1>
color                                
azul  -2.666667          2  -4.333333
negro  2.000000          3   2.000000
rojo  -3.500000          2   3.500000

It’s not advisable because: * It generates 2 levels of columns: it’s hard to “flatten” * It makes renaming the columns even harder. But at least it’s better than the previous method because you can use generic functions.

Suggested method

The method I recommend is agg but with explicit definition of the column names.

After the groupby, you apply the agg method by passing it an invented name for the desired parameter and giving it a tuple (column, function) as the value.

The generated columns immediately get the desired name!

import pandas as pd
df_g = df.groupby("color").agg(
        promedio_x = ("x", lambda _: _.mean()),
        conteo_y = ("y", lambda _: _.nunique()),
        promedio_y = ("y", lambda _: _.mean()),
).reset_index()
print(df_g)

   color  promedio_x  conteo_y  promedio_y
0   azul   -2.666667         2   -4.333333
1  negro    2.000000         3    2.000000
2   rojo   -3.500000         2    3.500000

The groupby can receive a list of columns in case you need a more detailed level of aggregation.

Note that the columns to be generated are passed as function parameters (promedio_y) and not strings ("promedio_y")!!!

Multicolumn

An interesting improvement is to use the above to work with multiple columns at the same time. Let’s look at a concrete example: calculating different metrics (distances) with respect to the origin (0,0).

To be able to apply the previous trick, we need to “pack the columns” into a variable

df["points"] = list(zip(df.x, df.y))
print(df)

   color  x  y    points
0  negro  1  1    (1, 1)
1  negro  2  2    (2, 2)
2  negro  3  3    (3, 3)
3   azul -2 -4  (-2, -4)
4   azul -5 -5  (-5, -5)
5   azul -1 -4  (-1, -4)
6   rojo -2  3   (-2, 3)
7   rojo -5  4   (-5, 4)

And now we can apply different functions, although in general it will be better to define them as functions (rather than lambda functions) to better express the relationships to use between the elements:

def norm_1(points):
    d = points.apply(lambda point: abs(point[0]) + abs(point[1]))
    return d.mean()

def norm_2(points):
    d = points.apply(lambda point: (point[0]**2 + point[1]**2)**.5)
    return d.mean()

def norm_inf(points):
    d = points.apply(lambda point: max(abs(point[0]), abs(point[1])))
    return d.mean()

df_agg = df.groupby("color").agg(
        norm_1_avg = ("points", norm_1),
        norm_2_avg = ("points", norm_2),
        norm_inf_avg = ("points", norm_inf),
).reset_index()
print(df_agg)

   color  norm_1_avg  norm_2_avg  norm_inf_avg
0   azul         7.0    5.222103      4.333333
1  negro         4.0    2.828427      2.000000
2   rojo         7.0    5.004338      4.000000

Non-Fiction: Visualization, Storytelling or both

In 2023 I realized that several of my interests come together in a single thread of “Storytelling”, both in the literary sense and in presentations.

• Made to stick, Chip and Dan Heath. A book that analyzes why some ideas are memorable and endure over time, while others are easily forgotten. Interesting.

• Bird by bird, by Anne Lamott. A book that mixes writing advice and autobiography, which I had been wanting to re-read. It talks about the difficulties of writing, and assorted advice for writing and life.

• Plotting and Writing Suspense Fiction, by Patricia Highsmith. On a podcast they recommended this book with the author’s perspective on writing.

• The DC Comics Guide to Writing Comics, by Dennis O’Neil. Recommended on some podcast, I read this guide which contains some interesting advice but isn’t essential either.

• Introduction to Data Visualization, by José Berengueres. One of the first books I read in relation to Data Storytelling/Visualization, and which is quite simple. I think the most relevant part is the advice and examples about minimizing the number of elements in a chart.

• Storytelling with Data, by Cole Nussbaumer Knaflic. The best book on Data Storytelling I have read so far. It has the technical background, examples, and is easy to read. A classic.

Fiction

• Un Mago de Terramar, by Ursula K. Le Guin. A little-known gem, written in the ’60s, that tells the story of a wizard’s apprentice and his struggle with a powerful adversary. I really liked how it talks about magic as knowing “the true name” of things. 100% recommended.
• La analfabeta que era un genio de los números, by Jonas Jonasson. The first novel I read by Jonasson, and it made me laugh a lot. Jonasson creates impossible and absurd situations, which are completely believable and plausible within the universe of his characters. In this case, a South African girl who ends up living in Sweden and who has an innate talent for mathematics.
• El viejo que saltó por la ventana y se largó, by Jonas Jonasson. It’s impossible not to love Allan Karlsson, a hundred-year-old man who decides to escape the nursing home where he lives. In this novel, we learn how he ends up stealing money from the mafia and also his wild stories around the world.
• El abuelo que volvió para salvar el mundo, by Jonas Jonasson. A sequel novel continuing the adventures of the same window-jumping old man, with a more up-to-date range of situations and political characters.
• Yomurí, by Cynthia Rimsky. A novel impossible to describe. Mapuche magical realism, perhaps? A group of people searching for the promised city for the Mapuche in southern Chile.
• Normal people, by Sally Rooney. I had seen that this book was quite popular at book fairs and in online sales, so I decided to give it a chance. An entertaining novel, but also sad and bittersweet. It talks about a couple’s relationship, their fights and bad decisions during their transition from school to university. I was surprised that the characters had a depth not usually found in these books.
• The incredible Mr Ripley, by Patricia Highsmith. Recommended on a podcast, I immersed myself in this classic thriller that I hadn’t read before. Very good.
• El jardín secreto, by Frances Hodgson Burnett. My son had to read this book for school, so I took the opportunity to read it to discuss it and to write a summary together. In the end, we asked chatGPT to write a summary, which was generally good although it added some hallucinations. The best part was that it prepared a list of possible questions about the book that were quite clever.
• *Silencio de los malditos, by Carlos Pinto. A novel by the creator of the ’90s television program Mea Culpa. The book was a bit predictable and played heavily on morbid curiosity, I didn’t find it very good.
• Ambos mueren al final, by Adam Silvera. I had also seen this book at many book fairs, so I gave it a chance. I found the main concept interesting, about being able to know 24 hours in advance when you’re going to die. Aside from that, a teen romance novel.
• El curioso incidente del perro a medianoche, by Mark Haddon. A novel written in the first person, where the protagonist is a teenager with autism. An exceptional book, because it makes you understand his perspective and experience his situations firsthand, and understand why he looks at and understands the world in a different way. Highly recommended. Here we ended up reading it as a whole family.
• Boom!, by Mark Haddon. Another book by Mark Haddon, a bit more for teens. It’s about some kids who discover that their teachers are aliens and a plan to conquer the world. Light but entertaining.
• La historia de tu vida, by Ted Chiang. A collection of disparate stories, but incredibly good. The degree of realism that Chiang achieves in his stories is amazing, because they are very strange or extraordinary situations, and yet the characters and their actions are completely believable.
• Trilogía de los Tres Cuerpos, by Cixin Liu. Made up of 3 books: El problema de los 3 cuerpos, el bosque oscuro, el fin del tiempo. It had been a long time since I read something that entertained me and made me think so much. I think each book surpasses the previous one.
• Rock, paper, scissors, by Alice Feeney. Someone recommended it on instagram as the crime novel with the biggest plot twist. Although it was entertaining, and had a good twist, it’s one of those novels that grips you until you finish it but you don’t really miss it afterwards.
• Sometimes I Lie, by Alice Feeney. I read a second book by Feeney, and the style and plot are very similar. Some twists even become predictable. Entertaining, but dispensable.
• El cuento de la criada, by Margaret Atwood. A dystopian classic that I hadn’t read and that surprised me a lot. 100% recommended. No spoilers.
• A la sombra del padre, by Hugo Kruger. An entertaining novel, one of those gems you sometimes get almost for free at the Viña del Mar book fair.
• El Faro, by Felipe Gonzalez. A novel set in Valparaíso, which narrates the adventures and misadventures of 2 cousins in the hills of Valparaíso.
• La ladrona de libros, by Markus Zusak. A novel about the Second World War, from the perspective of an orphan girl. It sounds sad and depressing, but the book has an ironic tone and is entertaining.
• Duna, by Frank Hebert. A science fiction classic, which I re-read for the second time after seeing the movie to remember some details.
• El visitante extranjero, by Julio Rojas. A period novel, about crimes that begin to happen in 1900s Valparaíso. Although it starts a bit slow, it has excellent period documentation, a good plot and interesting twists. One of my new favorite writers (fun fact: he’s the author of Caso 63, the podcast).
• Un mundo imposible, by Julio Rojas. A collection of short stories by Rojas. It has 2 of the best stories I have read, hugely recommended for readers in general, but especially for those passionate about technology or science fiction.
• El tigre, by Joël Dicker. A long short story or a very short novel. In any case, an interesting tale about a person who sets out to hunt a tiger to claim a reward. Very different from the crime novels with 16 plot twists characteristic of Dicker.
• Los Pilares De La Tierra, by Ken Follet. A novel about various characters from the Middle Ages, who cross paths in the city of Kingsbridge and its church. It has the ruthless, bloody and realistic tone of Game of Thrones, and great characters.
• El imperio final, by Brandon Sanderson. The first book of the “Nacidos de la bruna” trilogy. I had heard good things about this trilogy a while ago, and my son recommended I read it after he devoured the 3 books. A good fantasy novel, one that makes you wish allomancy really existed.x

Writing

I would like to: * Write a fiction book * Write a non-fiction book * Take part in some writing contests, particularly Santiago en 100 palabras (again) * Take part again in a writing workshop, such as the crime novel writing workshop

Learning and Creativity

I think I have a debt with my artistic skills: music and drawing. They are 2 dimensions I could explore a bit more. Additionally, learn to have a better sense of humor and laugh more at myself and my mistakes.

Sports

As a goal I want to do at least 5 minutes of exercise a day, Monday to Friday, throughout the year. I’ve discovered that it’s easier for me to meet the exercise goal at night, and that I can take advantage of the 7 minutes of waiting for the french press to do push-ups. I think that in addition to that, I need to go running if I want to be able to finish a marathon.

Technology

I want to keep participating in the Open Source space. I think here I don’t need to impose much discipline on myself, these are things that naturally emerge: * Increase my presence on twitter. Goal: 1000 followers. * Create new apps in Streamlit: complete the Machine Learning and Confusion Matrix saga, and make the datasaurus. * Participate in a free software development community, which could be the RISE one. * Attend the Latin American PyCons, particularly the ones Ariel is organizing in Salta. Maybe some other international one.

Reading

I have the goal of reading 10 non-fiction books and another 10 non-fiction books in the year 2022. For this I will set the goal of dedicating a little time every day and reading 1 fiction book and another non-fiction book per month. There are 2 months of tolerance for unfortunate events.

For the fiction books, I selected the ones Derek Sivers has marked 10/10: https://sive.rs/book

Topic: The world 1.- Open Borders - por Bryan Caplan

Topic: Writing 2.- On Writing Well - por William Zinsser

Topic: the mind 3.- Mindwise: How We Understand What Others Think, Believe, Feel, and Want - por Nicholas Epley 4.- The Time Paradox - por Philip Zimbardo y John Boyd

Topic: Learning 5.- The Listening Book - por W.A. Mathieu 6.- How Buildings Learn - por Stewart Brand 7.- E-Myth Revisited - por Michael Gerber

Topic: Life and death 8.- Forty Tales from the Afterlives - por David Eagleman 9.- The War of Art - por Steven Pressfield 10.- A Guide to the Good Life: The Ancient Art of Stoic Joy - por William Irvine

Other books: 1. Non designer design book 1. Resonate 1. Creativity INC 1. Invent and Wanter, Jeff Bezos 1. Zero to one 1. So good they cant ignore you 1. Numbers dont lie 1. Think Again 1. Antifragile 1. Almanak of Naval Ravikant 1. Why We Sleep: Unlocking the Power of Sleep and Dreams 1. The Sovereign Individual

Among the fiction books I would like to read or reread are: 1. Siddhartha, by Herman Hesse.