(a list, first)

1

There are definitions of success and happiness that don’t involve career. The last time I was around people who didn’t focus on their careers was Fleming County. But their lack of career ambition was primarily born from a lack of education and ambition.

2

There are people who devote their lives to caring for their section of the environment. They’re not just using it as a liberal talking point.

3

Similarly, being outside is a way of life. From intentional hobbies like climbing, biking, trail running, skiing, and rafting to just dedicating most weekends to camping with friends: being in nature is a top priority for some people.

4

To a different degree, being in nature on a regular basis is important to me too. Asma deciding she was going on her first ever real hike with me because I’m a “mountain person” is a funny way of learning this about myself. But I’ll be more intentional about getting out.

5

When sacrificing career (and therefore money) for passions, you can make significant happiness gains. But it does come with some amount of unescapable financial stress. I’m very grateful for finding some of my life’s genuine passion within the context of a high paying career.

6

There are important life lessons to be learned that may be difficult or impossible to learn within the confines of career/tech oriented cultures.

7

I’m most comfortable being goal oriented, and I’m okay with that. I can enjoy a flow festival, but the highlight might be helping a random food vendor with a broken cart setup their tent. I’ve dedicated energy to exploring the boundaries here. I’ll leave Tucson less goal-oriented than I entered it, but I can’t and have no desire to abandon my ambitions entirely.

8

You get to choose your own taboos. My version of Tucson doesn’t have many. Sometimes when I travel I have to reel it back in and remember that not everyone is comfortable talking about everything. A particularly beneficial topic that’s taboo in too many communities is talking about your struggles and weaknesses. Tucson’s a very non-pretentious group, so they don’t feel the need to hide or mask their shortcomings. Instead they can talk about, and therefore work on, things with their community.

9

I’m not infinitely flexible. That should be obvious, but I hadn’t internalized it until recently. Certain aspects of Tucson’s culture are viable, not qualitatively or quantitatively worse than other ways of living, but not for me.

10

The interplay between confidence, relationships and independence is convoluted and cyclical. Hypo-independent people (not an archetype I’ve seen around Tucson much) are overly attached which eventually rots their relationships. Hyper-independent people refuse to sacrifice their desires to accommodate others. At least, thats one reason for hyper-independence. Another has something to do with the fear of getting hurt, of being disappointed and let down by other people. Oddly, it seems like independent people are often viewed as confident. They’re just off doing their own thing! But it takes true self confidence to say:

“I trust in you. I understand I’m taking a risk that my feelings may be damaged. But, I believe it will be worth the risk. And even more importantly, I trust in my own capabilities and robustness. I will recover if this interaction or relationship goes south.”

Tucson is full of independent people. It’s also full of love. Having loved and lost genuinely is the preferable route.

11

Tucson has let me glimpse the purpose of poetry. I used to be silent until I could express my emotions accurately and precisely. There are situations that call for vocalizing your emotions with the words you have on tap. Paint the picture of your soul; the painting will convey what you need to convey better than words trying to photograph your brain under a proverbial microscope.

This is a draft I wrote in Summer, 2025. I’m posting about it now because I don’t want it to get lost. Maybe another version will come later.

Story time

Our school wasn’t super well taken care of.

One day during lunch I had to pee; because it was lunch period, there were a number of other guys in the bathroom and all the urinals were full. So I went into a stall and started peeing. As I was doing so, I started looking at what was going on in the toilet bowl. It was a busy situation.

It hadn’t been flushed in a while. There was tobacco juice because a lot of students secretly dipped, a decent bit of toilet paper, and a lot of urine. Thankfully nobody had taken a dump. But, underneath all of that was 80% of an uneaten apple that someone had casually munched on then tossed in the toilet. You know… as you do.

As I’m processing all of this, I hear a click.

We’ve got automatic toilets. Powerful ones. As soon as the suction starts, the apple immediately gets pulled down the drain until it perfectly stops up the toilet. Simultaneously, a super powerful jet of water comes out as part of the flush. Except, it doesn’t have anywhere to go because the apple has blocked it. Instead of a normal flush, the mix of tobacco juice, urine, and toilet water splash off the apple and form a wall of water that heads directly towards me.

Thankfully, I had just enough time to… open my mouth in shock. My pants are soaked, my shirt is soaked, my hair is wet. There’s piss in my mouth. I stumble out of the stall and everyone stares at me because this wall of water wasn’t confined to the stall. To some degree, it had sprayed the whole bathroom.

The first person to make a noise is Daniel. Because of course Daniel is there. Daniel was a year older than me, one of the captains of the football team, and surprisingly intelligent to boot. None of that would have matter if he wasn’t dating Hannah. Hannah was a cheerleader in my year who, while we had never formally dated, I had a “thing” with for a few years. So, Daniel sorta hated me. Daniel starts dying laughing.

The only thing that’s kinda dry on me is the back of my shirt. The only thing I can think to do is pull it off, and use the back to dry off my face a little. Because Daniel is laughing, all the other guys are laughing too.

Again, I don’t really know what else to do, so I wander out of the bathroom, still with no shirt on, to head to the front office. As soon as I step out, I run into four people. One is a grumpy teacher, who has probably come to see what the noise is about. Hannah is also right there. So are two of the other cutest girls in school. Turns out the cheerleaders have set up a fundraising table outside of the lunch room. They’re all staring wide-eyed.

The grumpy teacher immediately starts in, “What’s going on in there?!”. I haven’t found my words yet and mumble something like, “Uh, there was an apple in the toilet.”. Daniel rolls out of the bathroom still laughing. “And why don’t you have a shirt on?!”, Now I’m a little annoyed because the answer to that question should be obvious. “Because… it’s covered in urine?” The girls are giggling, and the teacher’s still confused, “Well… you need to go to the principle’s office right now!” Fully frustrated at this point, I snap “Yep, that’s where I was headed anyway.”

I make eye contact with Hannah, and wander off down the hall to sit in the front office for an hour before my mom can bring me dry close. They didn’t even let me go home to shower, I had to finish out the rest of the day.

Downloading the data is one thing, but digging through it is another. I decided that now was as good of a time as any to try my first coding agent. So I know have codex on my computer, and I gotta tell ya… the first experience was really, really nice. I didn’t do a ton with it. But I had it setup a repo managed by uv, tossed it the path to all my data and told it to write a tool to find the 10 longest messages I’ve ever sent. I think the only error it produced was forgetting to run

uv sync

before trying out the command.

I wasn’t sure if a heuristic as simple as “longest sent message” was going to be enough to uncover any stories. It’s not like I usually write an entire tale in one message. And I was mostly correct. Of the 10 messages, only one of them contained a story and only then because I’d wholesale copied/pasted a back-and-forth from one conversation to another.

In a later post I’ll link out to the code in case anyone else wants to download their data too. In the meantime, here’s a story from high school, roughly as told in 2010:

Poo Shoe

Freshman year of high school, we had a class called PAWs: every month we went to a random teacher and they taught us something the teacher thought was interesting but wouldn’t fit in a normal course. One time, my PAWs class was in the school barn. Yes, our school had a barn. The barn wasn’t exactly my favorite place to spend school hours, so I quietly made it through the class, luckily without incident.

I got back up to the school and next period was English class. I made it a quarter of the way through the class before people started making fart jokes. Halfway through the class I realized it was me and that I must have stepped in dog crap or something during PAWs. I didn’t want to continue stinking up the class, so I asked the teacher if I could take off my shoe and put it outside the door. She said yes, so I did.

At the end of class I went to get my shoe and it was gone. I asked around to the students hovering around waiting to get into their next period, but everyone denied knowing anything about it. Someone had stolen my shoe and was lying about it! There really wasn’t anything I could do, and my next period was starting, so I had to just hobble down to my next class and pray that I wouldn’t get asked too many questions.

I had to go to class with only one shoe and try to explain what happened. The next block, my teacher told me to go down to the office and see if anyone had turned it in or something. I told the lady in the office someone had stolen my shoe, and she was basically like, “I don’t care, go back to class. There’s nothing I can do about it.” So I said okay and went back to class, and I spent the next two blocks without a shoe.

Then I went to my last class and we had this strange sub. She realized I didn’t have a shoe and questioned me. I wasn’t really that mad about the incident, but when I told her what happened she completely freaked out: “NO CHILD SHOULD HAVE TO GO THROUGH THIS KIND OF HUMILIATION!!!” “WHAT IS WRONG WITH PEOPLE AT THIS SCHOOL???” “NOBODY SHOULD HAVE TO GO THROUGH SOMETHING LIKE THIS!! IT’S JUST RIDICULOUS!!!”

We had an assembly, so the sub was free to do what she wanted because all the students were in the gym. She almost literally grabbed me by the collar and took me around the school, looking in random places for my shoe. The entire time she kept chanting about how no student should have to suffer through this kind of humiliation. She was taking the lid off trashcans and digging around in the trash looking for my shoe.

And then, all of a sudden, in one of the trashcans, there was my shoe. This was about 10 minutes before it was time to go home, so I just scraped as much of the poo off my shoe as possible, and got on the bus to go home.

The slides are probably a little hard to get value from on their own, but there’s a little more content in the speaker notes.

Maybe one day I’ll write these thoughts up more completely.

I’ve been asking this question to people for nearly a decade now (ever since I watched the Black Mirror episode “San Junipero.” If you haven’t seen it, it’s worth a watch—we’ll loop back to it at the end of this essay). For now I’ll just say that it’s fitting timing to write this since Season 7 of Black Mirror was released today. I have high hopes!

Anyway, I’ve asked this longevity question hundreds of times to a very wide range of people. It’s a fallback question I keep in my pocket for when there’s more time than topics. Even though I haven’t kept formal statistics, I’ve had the conversation enough to pull out themes and frequencies. Here’s how the conversation typically breaks down:

1. Most people ask some clarifying questions
2. They pick an age. Although I didn’t tell them this until later, there are only 6‑ish meaningfully different answers I’ve heard people give.
3. We talk about their rationale and what I’ve gathered from others.

Clarifying questions

People usually want a little more info on the mechanism of the hypothetical. How’s their health? What’s the aging mechanism? I tell them they can have a choice:

• Proportional aging. If they want to live to be 1,000, they would spend ~60 years in their “60s.”
• Freeze at a single “perfect” age. For example, they could keep the health of a 25‑year‑old indefinitely, then very rapidly decline in their final years.

The second most common question is: “Do my loved ones also live the length I choose?”

I’ve always answered this with a:

“No, they live a normal human life span.”
It’s how I answered the first time I was asked, and I’ve just stuck with it. Writing this down, however, makes me want to explore how this severe penalty affects people’s answers.

These are by far the two most common clarifications. Other questions include how memories work, more details on sickness, whether other people know, etc.

Answers

The most common answer I get to this question is “I want to live a normal human lifespan.” Numerically, people tend to express this in the 80–100 yr range. Often, people who respond with this answer mention significant hesitations about outliving friends and family.

Nearly equal in frequency is “slightly longer than average,” which—if we take the numbers people use (100–120)—actually means something more like “in the top 1 %. of human lifespans” These people are comfortable with their lives and want to keep doing it a little longer than they expect to. But they usually seem wary of trying to imagine a qualitatively different human experience.

Next up is infinite. Some people—the wild ones—want to live forever. They want to have it all. Frankly, I suspect this is the worst possible answer, for reasons I’ll explain later.

Now we’re entering the uncommon territory. Even though this answer is less common than the previous two, living 2 × (160–200 years) the normal human lifespan is surprisingly common. People seem to answer this way when they’re thinking along the lines of “This life thing is pretty great. I’d do it again.”

The second‑rarest answer people routinely give is “in the 50s or 60s.” Some portion of people misunderstand the hypothetical and think they’d have to spend eternity in poor health. Even if I clarify the prompt, they double down on their original answer. But others—often people who have lived difficult lives—understand the prompt but want no more to do with life than is strictly required.

And finally, the least common answer I get is the “really long, non‑infinite time.” This is anywhere in the 1000 to 1,000,000 range.

Conclusion

It’s maybe obvious by now—or maybe it was obvious at the beginning—that a person’s answer to this question serves as a proxy for how much someone is enjoying their life. I’m simplifying, but there’s a strong correlation. The other (anti)-correlation I’ve noticed is that the older the person answering is, the shorter their answer is going to be. I don’t think I’ve ever asked this question to someone in their 60’s and had them say, “I wanna live for a billion years”. People seem to naturally find acceptance of death as they age.

Immortality

There’s also the matter of the people who want to live forever…

So many people talk about how “San Junipero” is their favorite episode of Black Mirror. It’s the one with the happy ending—the two lovers riding off into the virtual sunset to spend eternity with each other. The problem with this popular fairytale interpretation is that it neglects a thoughtful inclusion of the “Quagmire” scene. In the middle of the episode, the main character goes to a club called The Quagmire. The rave lit warehouse is filled with people having a seemingly fun (if extreme) time, but the conversational backdrop is quite explicit about the truth: The Quagmire is home to all of the system’s oldest residents. The ones who have already lived several normal human lifespans’ worth of experiences.

According to the conversation, there’s broad consensus that with each passing year, human brains must go to more extreme lengths to consider something novel, to feel alive and passionate, to have a reason and a purpose. When I see the pair of lovers drive off into the sunset in the final scene, I adopt the same conclusions as the Black Mirror writers: the lovers will eventually end up at The Quagmire, along with most or all of San Junipero’s residents.

And they’re equally doomed to stop going to The Quagmire one day.

Neither going to The Quagmire for the first time or leaving it for the last time are moral failings, they’re simply a side effect of humans propensities towards generalization while simultaneously craving novelty. As every drug user knows, “you only get your first time once”. But this tendency towards normalization applies to everything, not just drugs and other vices. I remember the first time I saw a Saguaro cactus. It was so cool that I stopped and took a picture of it. Now I could see hundreds in a day and not even register them.

One could argue that fizzling out, getting depressed, and eventually committing suicide is a small price to pay for immortality. I’m not so sure. Our biology simultaneously compels us toward self‑preservation just as strongly as it demands purpose.

The suicide quagmire might drag on for quite some time, but not forever…

Counter action

Again, we don’t have to be faced with possible immortality for these issues to be relevant. We all face this question of, “How do we keep our spark for life?”, or the equally cliche “How do you keep your sense of childlike wonder?” They’re cliches because they’re universal questions.

My answer is similar to the conclusion of John Salvatier’s blog post, Reality has a Surprising Amount of Detail:

This problem is not easy to fix, but it’s not impossible either. I’ve mostly fixed it for myself. The direction for improvement is clear: seek detail you would not normally notice about the world. When you go for a walk, notice the unexpected detail in a flower or what the seams in the road imply about how the road was built. When you talk to someone who is smart but just seems so wrong, figure out what details seem important to them and why. In your work, notice how that meeting actually wouldn’t have accomplished much if Sarah hadn’t pointed out that one thing. As you learn, notice which details actually change how you think.

I’d like to add an addendum: as we age, try to look for and appreciate finer and finer grained detail. Reality is like a fractal: as you look closer and closer, you’ll notice more and more details that weren’t perceivable from a higher-level perspective.

The trick is realizing that the value of perception is scale invariant.

There’s no more inherent value in noticing a forest than there is a flower in that forest.

Interview Question

I have a fasta file that has the following format:

>SEQ1
AGTCTGCTACGCTAGC
>SEQ2
GGTGCGATCGATCGTG
>SEQ3
TTCCCATGCTACGTCG
>SEQ4
TCAAACGCATCGCGAC

You can assume that:

• the fasta file is well formatted
• there is only one sequence line per header line
• each sequence in the file is the same length

Your task is to replace the pass portions of this code with functioning code, according to the name and docstring:

def read_fasta(input_fasta_path):
    """Given the location of a fasta file, return a list of sequences"""
    pass


def are_similar(seq1, seq2, threshold=3):
    "Given two sequences, using a Hamming distance metric to return a bool indicating whether or not the two sequences are similar"
    pass


def build_graph(sequences):
    """Given a list of sequences, return a graph representation of sequences that are similar via Hamming distance"""
    pass


def are_second_degree_neighbors(graph, seq1, seq2):
    """Given a graph and two sequences, return a bool indicating if they're second degree neighbors""" 
    pass


def main():
   pass

Please create any tests and data you need in order to feel confident that the code you write is correct. Similarly, structure main in any way you see fit to demonstrate correctness of the code.

Lastly, imagine we want this code to go into production. Please attempt to make coding decisions that produce production-level code.

ChatGPT-o1

(Thought for 29s)

#!/usr/bin/env python3
"""
FASTA Similarity Checker

This module provides functionality to:
1. Read sequences from a FASTA file.
2. Compute a Hamming-based similarity check between two sequences.
3. Build a graph (adjacency list) of sequences considered "similar."
4. Determine whether two sequences are second-degree neighbors within that graph.

Author: Your Name
Date: 2025-01-31
"""

from itertools import combinations
import sys


def hamming_distance(seq1: str, seq2: str) -> int:
    """
    Compute the Hamming distance between two equal-length sequences.
    
    :param seq1: First sequence (string).
    :param seq2: Second sequence (string).
    :return: The Hamming distance (integer).
    :raises ValueError: If the sequences are not the same length.
    """
    if len(seq1) != len(seq2):
        raise ValueError("Sequences must be of equal length to compute Hamming distance.")
    return sum(ch1 != ch2 for ch1, ch2 in zip(seq1, seq2))


def read_fasta(input_fasta_path: str) -> list[str]:
    """
    Given the location of a fasta file, return a list of sequences.
    Assumptions:
      - The FASTA file is well-formatted.
      - There is only one sequence line per header line.
      - Each sequence is of the same length.
    
    :param input_fasta_path: Path to input FASTA file.
    :return: A list of sequences (strings).
    """
    sequences = []
    
    with open(input_fasta_path, "r") as f:
        for line in f:
            line = line.strip()
            if not line:
                continue
            # Skip headers that start with '>'
            if line.startswith(">"):
                continue
            sequences.append(line)

    return sequences


def are_similar(seq1: str, seq2: str, threshold: int = 3) -> bool:
    """
    Given two sequences, use a Hamming distance metric to return a bool
    indicating whether the two sequences are 'similar.'
    
    :param seq1: First sequence (string).
    :param seq2: Second sequence (string).
    :param threshold: Maximum Hamming distance for which sequences are considered similar.
    :return: True if Hamming distance <= threshold, False otherwise.
    """
    dist = hamming_distance(seq1, seq2)
    return dist <= threshold


def build_graph(sequences: list[str], threshold: int = 3) -> dict[str, set[str]]:
    """
    Given a list of sequences, return a graph (adjacency list) of sequences
    that are similar with respect to a given Hamming distance threshold.
    
    :param sequences: A list of sequences (strings).
    :param threshold: Maximum Hamming distance allowed to declare two sequences connected.
    :return: A dict where keys are sequences and values are sets of adjacent (similar) sequences.
    """
    graph = {seq: set() for seq in sequences}
    
    for seq1, seq2 in combinations(sequences, 2):
        if are_similar(seq1, seq2, threshold):
            graph[seq1].add(seq2)
            graph[seq2].add(seq1)

    return graph


def are_second_degree_neighbors(graph: dict[str, set[str]], seq1: str, seq2: str) -> bool:
    """
    Given a graph (as returned by build_graph) and two sequences, return a bool
    indicating if they are second-degree neighbors.
    
    Here, "second-degree neighbors" means that seq1 is not directly connected to seq2,
    but there exists some intermediate neighbor 'n' such that seq1 -> n -> seq2 is a path.
    
    :param graph: The adjacency list (dict of sets) of sequences.
    :param seq1: The first sequence (string).
    :param seq2: The second sequence (string).
    :return: True if seq2 can be reached from seq1 by exactly two "hops." 
    """
    # If either sequence is not in the graph, cannot be neighbors
    if seq1 not in graph or seq2 not in graph:
        return False
    
    # If they're first-degree neighbors, they're not strictly second-degree neighbors
    if seq2 in graph[seq1]:
        return False

    # Check for an intermediate neighbor
    for neighbor in graph[seq1]:
        if seq2 in graph[neighbor]:
            return True
    
    return False


def test_all():
    """
    A simple test suite to ensure correctness of the code.
    """
    # Test Hamming distance
    assert hamming_distance("AAAA", "AAAT") == 1
    assert hamming_distance("AAAA", "AAAA") == 0
    try:
        _ = hamming_distance("AAA", "AAAA")
    except ValueError:
        pass  # expected

    # Test are_similar
    assert are_similar("AAAA", "AAAT", threshold=1) is True
    assert are_similar("AAAA", "AATT", threshold=1) is False
    
    # Build a small graph and test adjacency
    seqs = ["AAAA", "AAAT", "AATT", "TTTT"]
    g = build_graph(seqs, threshold=1)
    # neighbors of "AAAA" should only be "AAAT"
    assert g["AAAA"] == {"AAAT"}
    # neighbors of "AAAT" should be "AAAA" and "AATT"
    assert g["AAAT"] == {"AAAA", "AATT"}
    
    # Test are_second_degree_neighbors
    # "AAAA" -> "AAAT" -> "AATT" => second-degree neighbors
    assert are_second_degree_neighbors(g, "AAAA", "AATT") is True
    # But "AAAA" and "TTTT" are not even close (distance=4 each from "AAAA")
    assert are_second_degree_neighbors(g, "AAAA", "TTTT") is False
    
    print("All tests passed!")


def main():
    """
    Demonstration of how to use the above functions in practice.
    """
    # 1. Run tests to ensure code works correctly
    test_all()
    
    # 2. Example usage with a sample FASTA file (replace 'example.fasta' with a real file path)
    fasta_file = "example.fasta"

    try:
        sequences = read_fasta(fasta_file)
    except FileNotFoundError:
        print(f"Error: FASTA file '{fasta_file}' not found.")
        sys.exit(1)
    
    print("Sequences read from FASTA file:")
    for s in sequences:
        print(s)
    
    # 3. Build the similarity graph
    graph = build_graph(sequences, threshold=3)
    
    # 4. Demonstrate neighbor checks (customize seq1, seq2 to your input)
    if len(sequences) >= 2:
        seq1 = sequences[0]
        seq2 = sequences[1]
        print(f"\nChecking if '{seq1}' and '{seq2}' are similar (threshold=3): {are_similar(seq1, seq2, 3)}")
        print(f"Are '{seq1}' and '{seq2}' second-degree neighbors? {are_second_degree_neighbors(graph, seq1, seq2)}")
    else:
        print("Not enough sequences to demonstrate neighbor checks.")


if __name__ == "__main__":
    main()

Running the script

I used the following test file as example.fasta:

>SEQ1
AAAAAAATTTTTTTCCCCCCGGGGGGG
>SEQ2
AAAAAAATTTTTTTCCCCCCGGGTTTT
>SEQ3
GGGGGGGTTTTTTTCCCCCCGGGGGGG
>SEQ4
AAAAAAATTTTTTTCCCCCCGGGGGTT
>SEQ5
AAAAAAATTTTTTTCCCCCCGGGGGGC

I’d expect the o1 script to say that the first two sequences are neighbors.

❯ python chatGPT_o1.py
All tests passed!
Sequences read from FASTA file:
AAAAAAATTTTTTTCCCCCCGGGGGGG
AAAAAAATTTTTTTCCCCCCGGGTTTT
GGGGGGGTTTTTTTCCCCCCGGGGGGG
AAAAAAATTTTTTTCCCCCCGGGGGTT
AAAAAAATTTTTTTCCCCCCGGGGGGC

Checking if 'AAAAAAATTTTTTTCCCCCCGGGGGGG' and 'AAAAAAATTTTTTTCCCCCCGGGTTTT' are similar (threshold=3): False
Are 'AAAAAAATTTTTTTCCCCCCGGGGGGG' and 'AAAAAAATTTTTTTCCCCCCGGGTTTT' second-degree neighbors? True

Deepseek R1

(Thought for 176s)

from collections import defaultdict


def read_fasta(input_fasta_path):
    """Given the location of a fasta file, return a list of sequences"""
    sequences = []
    with open(input_fasta_path, "r") as f:
        for line in f:
            if line.startswith(">"):
                # Read the next line as the sequence
                seq = next(f).strip()
                sequences.append(seq)
    return sequences


def are_similar(seq1, seq2, threshold=3):
    """Given two sequences, return True if their Hamming distance is <= threshold."""
    if len(seq1) != len(seq2):
        return False
    hamming_distance = sum(c1 != c2 for c1, c2 in zip(seq1, seq2))
    return hamming_distance <= threshold


def build_graph(sequences):
    """Given a list of sequences, return a graph where keys are sequences and values are lists of similar sequences."""
    graph = defaultdict(list)
    n = len(sequences)
    for i in range(n):
        for j in range(i + 1, n):
            if are_similar(sequences[i], sequences[j]):
                graph[sequences[i]].append(sequences[j])
                graph[sequences[j]].append(sequences[i])
    return graph


def are_second_degree_neighbors(graph, seq1, seq2):
    """Return True if seq1 and seq2 are connected via exactly two edges in the graph."""
    if seq1 == seq2:
        return False  # A sequence cannot be a neighbor of itself in this context

    # Check if they are direct neighbors
    if seq2 in graph.get(seq1, []):
        return False

    # Check if there's any common neighbor that connects them
    for neighbor in graph.get(seq1, []):
        if seq2 in graph.get(neighbor, []):
            return True
    return False


def main():
    # Example test case with a small FASTA file
    test_fasta = [">SEQ1", "AAAA", ">SEQ2", "AAAT", ">SEQ3", "AATT", ">SEQ4", "ATTT"]
    # Writing test data to a temporary file
    with open("test.fasta", "w") as f:
        f.write("\n".join(test_fasta))

    # Read the sequences
    sequences = read_fasta("test.fasta")
    print(f"Read {len(sequences)} sequences: {sequences}")

    # Build the graph
    graph = build_graph(sequences)
    print("Graph edges:")
    for seq, neighbors in graph.items():
        print(f"{seq}: {neighbors}")

    # Test are_similar
    print("\nTesting are_similar:")
    print("SEQ1 and SEQ2 (threshold=1):", are_similar(sequences[0], sequences[1], 1))  # True
    print("SEQ1 and SEQ3 (threshold=1):", are_similar(sequences[0], sequences[2], 1))  # False

    # Test are_second_degree_neighbors
    print("\nTesting are_second_degree_neighbors:")
    print("SEQ1 and SEQ3:", are_second_degree_neighbors(graph, sequences[0], sequences[2]))  # True
    print("SEQ1 and SEQ4:", are_second_degree_neighbors(graph, sequences[0], sequences[3]))  # False


if __name__ == "__main__":
    main()

It’s clever that it chooses to write out then read in a file in main.

If we just run the script, we get the output:

❯ python deepseek_R1.py
Read 4 sequences: ['AAAA', 'AAAT', 'AATT', 'ATTT']
Graph edges:
AAAA: ['AAAT', 'AATT', 'ATTT']
AAAT: ['AAAA', 'AATT', 'ATTT']
AATT: ['AAAA', 'AAAT', 'ATTT']
ATTT: ['AAAA', 'AAAT', 'AATT']

Testing are_similar:
SEQ1 and SEQ2 (threshold=1): True
SEQ1 and SEQ3 (threshold=1): False

Testing are_second_degree_neighbors:
SEQ1 and SEQ3: False
SEQ1 and SEQ4: False

Conclusions

• Both of these models pass the test. By comparison, probably 1/3 people I interviewed would pass. Many bioinformaticians get bogged down in even simple data structure / algorithm style questions.
• There’s a little room for ambiquity in the question. I used to spend a lot of time in the interviews answering questions and clarifying that ambiquity. Both models make reasonable choices.
• I like to see both itertools and collections utilized, as a demonstration of knowledge of the standard library. It’s interesting that each model uses one but not the other.
• o1’s tests were a little more robust, IMO.
• R1 took 6x the time to solve the problem.
• I have a slight qualitative preference for o1’s code.

// TODO Comment on the set approach for 2nd degree neighbors

And finally:

• The next time I interview a BFX role, I should come up with another question. This one is “solved”.

In the end, my trips were a bit more clumped than I had hoped. But, I still had a lot of adventure this year. From beautiful nature, to joyful times with friends, it’s easy to find reasons for gratitude in the past year.

1. Sedona Round 1 (February)

My first trip of the year had some twists. Twists like having RSV or something the whole trip, or falling on an agave plant and literally nearly losing an eye. But it’s hard to overstate how beautiful the color juxtapositions of the red rocks, evergreens, and pure white snow was.

2. Vegas (February)

This night in Vegas seeing a Benny Benassi show contained one of the most memorable conversations of the year. I was at a table with a group of neurosurgeons who had covered for the wild evening. I was in a delicate headspace because this was before we’d started consulting, so I was feeling a lot of financial pressure and a tinge of, let’s admit it, masculine fragility.

Halfway through the night, well past sobriety, one of the doctors pulls me over and shows me his phone. “Check this out. It’s my list of startup ideas. It’s so cool that you get to just follow your dreams and start your own company.” I quipped back, “Hey, all you have to do is go do it. You could start one!” He just grunted. “No I can’t; I’ve got golden handcuffs. I’d lose my house, my girlfriend would leave me. I’ve gotta pay alimony. You’re free.”

I’d spent the whole night envious of these guys throwing money around. You know, they’d really made it in life. And yet…

3. Kentucky Round 1 (June)

You know your family is cool when you can text “Hey, I’m coming into town” three days before you show up and there’s still an impromptu cookout for you with 20 people (and three more on the way).

The trip to KY was also a great time to catch up with the nuclear family and spend too much time playing Stronghold Crusader with Mahlon.

4. North Carolina (June)

A night out on the town in Durham. That evening was a good reminder that time’s flying by, that everyone is changing (in good ways!), and that it’s absolutely worth finding the time to catch up with old friends.

Another highlight of this time was getting to meet Wes’ and Sherif’s kids for the first time. Energy for days.

5. New York City (June)

Speaking of catching up with old friends, I did a lot of that in NYC too. But, this was my first time getting to wander around NYC at night by myself, and I was surprised by how pleasant it was! My baseline is SF. At least in my initial impressions, there’s no comparison. I saw more cops in NYC than I did homeless people.

But… SF is prettier, if sleepier.

6. Sedona Round 2 (October)

Even though this was an out of town trip for a wedding, this photo is a great representation of the appreciation I’ve felt this year for my Tucson crew. And even though this is a boys pic, the women in my Tucson life are equally amazing humans.

7. Denver (November)

This is a goofy representation of my third year at ASHG. I took it a bit more chill this year. I didn’t even lose my voice. I was hoping that I’d be inspired to write up an entire post on ASHG. But between the sizable amount of time I spent catching up with people in 1:1 conversations and the seeming lack of significant innovation this year, there wasn’t much I felt like talking about.

8. Gaithersburg (November)

Snack fridge…

9. Grand Canyon (November)

When your Canadian friend says she wants to see the Grand Canyon, you make it happen. And you sing road trip karaoke the whole way there.

10. Cochise (November)

This photo has to be dedicated to community and the work of community building. Community has been a reoccurring theme this year.

While this picture was taken on Thanksgiving day, it resonates with my Thankful this year during Friendsgiving. I wrote up a few notes to myself before opening the circle. Here they are:

• Every year, I get it easy.
• I get to be the first to say, “I’m thankful for friends for my community!” And then you all have to figure out something to say so it doesn’t sound like you’re a copy cat.
• But also, some of you have been hearing me say versions of this for 10 years now. So I gotta keep it fresh.
• It’s been a heady, existential year. I’ve had so many conversations about our purpose in life. About self worth and confidence. About what it means to be successful. About what it means to fail.
• I’m sure I’m not alone in this headspace. Especially for those of us who are more goal oriented, it’s tempting to measure life by progress towards those goals and as a side effect it’s also tempting to form this hierarchy and rank yourself and others relative to these goals.
• I’m so glad I’m in a community that’s been able to instill in me that life is more complex than that linear perspective.

• This year, like always, I’m thankful for each of you. But more specifically, this year, I really appreciate the aspects of each of you that make you unique and weird individuals. And I hope you appreciate those things about yourself too.

• People who take night hikes to look for bugs
• People who host and attend puberty themed birthday parties

• People who do group photo shoots of rainbows in their underwear

• So again, what I’m truly thankful for is each one of you and the uniqueness you contribute to the world. Thanks you for existing.

11. San Francisco (December)

Every time I visit San Francisco, I see something new. Every time I visit San Francisco, it feels more like home than any other city I’ve visited.

12. Kentucky Round 2 (December)

Speaking of home, there’s nothing like visiting secret little spots you first discovered when you were 10. These places trigger an acute nostalgia that call into question your understanding of time.

Humble is as Humble does.

Subtle can be Muscle flush.

Stumble now, and trouble thus:

Jumble, Mumble, come Fuss.

Puzzle is the Shuffle of;

Knuckle thrown, Shovel Brush.

Huddle close and Rubble shove.

Buckle not, Chuckle, find Hush.

Cuddle, cuddle, cuddle, love.

- Mahlon Kirk, October 1st, 2024

1. Melons. I think I’ve mostly been exposed to old melons that have sat precut for too long and have gone mushy. There’s also a high likelihood that they’re surrounded by other fruit and taste like an undistinguishable blend. I’ve probably bought a half dozen melons in the past year (coming from having never bought one). They’re definitely going to stick in the rotation.
2. Smoked salmon. A little bit pricy, but so amazing. I could legit eat pounds of smoked salmon.

This week, I picked up a new fruit from Costco, a rambutan.

They’re pretty similar to lychee, which I’ve had once before.

You cut them in half (the seed isn’t too bad to cut through):

And the fruit easily and cleanly peels away from both the outer shell and the pit. One little layer of the pit stays attached, but I just ate that. It adds a little bit of crunch, but in a pleasant way.

The texture is great. The flavor’s a little too mild to be considered great but is still pretty good.

But this year I made one simple goal:

Read or listen to 12 books this year

Reading has been one of my favorite activities since childhood, and I’ve always found it worthwhile to carve out some time for books. This year, I made it through one a month, with a bonus collection of scifi and fantasy short stories to bring the total to 13! Here’s a link to the full list. Unfortunately, I didn’t make it through Kafka on the Shore before it needed to be returned, and I haven’t had a chance to renew it.

The list of topics is super diverse, and if it seems a little random, that’s because it more or less was. My process for choosing audiobooks this year was to open the Libby app (accessed through my local library card), filter for currently available audiobooks, sort by most popular, and scroll until something caught my eye. Something that surprised me, however, is that half the books are (auto-)biographies:

1. Permanent Record (Edward Snowden)
2. Yes, Please (Amy Poehler)
3. Brain on Fire (Susannah Cahalan)
4. Kitchen Confidential (Anthony Bourdain)
5. Surrender (Bono)
6. The Splendid and the Vile (Winston Churchill)

That’s equal to the number of biographies I read from 2014 to 2022. But I loved diving into these people’s lives and was consistently surprised at how connected I felt to the content and messages of these books, regardless of how distant the individual’s existence seemed from my own reality. Kitchen Confidential and Surrender, in particular had surprising overlap with many of my emotions around starting Lodestar this year. It turns out that Bourdain and I have similar philosophies on hiring a team. And Bono could teach me a lot about sales.

Erin asked me if I had a favorite, but I don’t. Apples to oranges and all that. I will say that, despite my rule about not starting series that haven’t been finished, I would absolutely recommend The Way of Kings to anyone, and am looking forward to getting my hands on the next book. It’s impressive to compare The Way of Kings to Elatris, which I read last year and is Sanderson’s first book. The amount of growth he had as an author between the two books is phenomenal.

The final observation I want to make is that reading each of these titles now makes me recall at least one specific instance of listening to the book outside the house. Whether it was walking in San Francisco, in the airport, in the neighborhood behind my house, or while lifting at the gym, I have an association between the book and other things that were going on in my life at the time. I can’t prove this, but a hypothesis I have is that it’s been harder for me to remember work related this I’ve learned since WFM started during covid. So many of the conversations I’ve had and things I’ve learned recently are only anchored to the same wall or corner of my room. Pre-covid those same memories would be tied to someone’s 3D full-resolution facial expression, or a bar smell, or standing awkwardly over someone’s shoulder while they code.