Volunteering has taken a few different forms for me. I’ve been occasionally having career-related conversations with students and recent alumni of my college, many of whom weren’t born during the dot-com bubble that burst while I was in school, and who need to be creative and unorthodox in their job searches. I’ve also been volunteering my time for housing, public transportation, and walkability of my local neighborhoods. The recent news that Sound Transit has delayed its voter-approved extension of light rail to Ballard led to the creation of Save Ballard Rail, an advocacy group that I’ve joined. Our group already has the attention of transit officials, and we’re going to keep making noise until we get the project done sooner. Earlier this year, I also answered a call to become an SEA Ambassador, a part-time volunteer assisting travelers at Seattle–Tacoma International Airport, in anticipation of a surge of travelers for the FIFA World Cup.
From my experience, volunteering can take as much time and energy as I’m willing to put in. Most of the organizations I volunteer with have regular meetings, ongoing projects, and opportunities to join larger efforts if I want to. Some organizations that lobby government officials have meetings with those officials, or with other organizations with similar goals, for example. After having done a few exciting and exhausting shifts at the airport, I’ve been exploring its full-time volunteer program, which requires a background check and a six-month commitment. One of my fellow volunteers, an employee at the Port of Seattle, learned about my professional background and suggested that I join the port as an employee to improve its software; I’m considering that, too.
It’s really inspiring to see how volunteer-driven campaigns and organizations attract people who have a common purpose. Particularly when my fellow volunteers live in my same neighborhood, I feel like I’m making more meaningful connections through volunteering than I had been making through work. It feels genuinely better to be doing work, even without pay, that helps people out. It’s possible to find professional work with a social good, but at a time when most tech companies that I talk to seem obsessed with novel, trendy, high-margin, high-growth technologies, I find myself drawn to simpler and more pragmatic projects.
Sometimes, I get questions about whether I consider myself “retired.” I’m still avoiding that word, and at the same time, I have the privilege to be very selective about where I’d consider working next. My experience as a volunteer has made me wonder whether there’s paid work that aligns with my interests in improving my city and helping the people in and around it. That work is likely to pay less than the market rate for engineering in artificial intelligence or other hot fields, but that’s less of a concern for me. I’m also interested in work on open source projects that can be used by multiple organizations, or as a consultant to help implement or manage operations for a nonprofit or social-good organization, even if it’s part time. Managing my time well also includes taking time off for recreation and improvement, especially during a beautiful Washington summer. Just because I’m not working doesn’t mean my calendar is empty; to the contrary, it can get uncomfortably full if I’m not careful. I’m seeking good uses of my time that I can sustain for a while.
]]>My integration into my new city started with a realization that I’d have to change my speech patterns a bit. Shortly after moving to Seattle, I was having lunch at work with my teammates, and I referred to The Oregon Trail, pronouncing “Oregon” in my New York accent, rhyming it with “octagon”. The conversation ground to a halt, and the rest of the table fixated on my bizarre pronunciation. If I was going to make Seattle my home, I realized that I’d have to learn to speak like a Pacific Northwesterner. For example, I now usually rhyme the first syllable of “orange” with “boar,” not “car.” I can pronounce “coffee” more like a dictionary’s pronunciation guide, and less like Linda Richman, the stereotypical New York housewife who hosted “Coffee Talk” sketches on Saturday Night Live. (In practice, Seattleites don’t often use that word, anyway; drip-brewed coffee is called “drip”, and straight espresso is either a “doppio” or “spro.”) The Yiddish-derived slang words common in New York speech, like “schmutz” or “schlep”, mostly effect bemused smiles in Seattle. When visiting New York, I try to code-switch back into the accent I grew up speaking, not always easily; at dinner in Manhattan last winter, I forgot whether I should pronounce “Mario” with its first two syllables like “marry,” in the style of New Yorkers like Mario Perillo, “Mr. Italy”, or in the modern Italian way, as most Nintendo fans in the U.S. and Japan say. Although my family insists that my accent has been neutralized, I’ve been recognized as a New Yorker within a minute of having met someone raised here, whether because of my aggressiveness, my sarcasm, or my revealing and grating pronunciations of words like “class” and “last.” I have a sixth sense for “New York energy” myself, identifying a few ex-residents of the tri-state area by their behavior.
I like my adopted home, and I’m also an advocate for improving it. I became interested in “urbanist” activism, in favor of dense housing, mass transit, and pedestrian and bike accessibility, when I started going to public information sessions about Capitol Hill station, which began construction in 2008 and opened in 2016. After a decade of ignoring signature gatherers for ballot initiatives, I volunteered as one, being part of two successful campaigns to get Seattle’s social housing developer created and funded, with a payroll tax on companies that have very well-paid employees in the city. During my career break, I’m showing up at events where I can advocate for housing and transit during the typical business week. Seattle continues to grow in population and prosperity, and the city needs more housing and transit to keep the cost of living from rising even more than it has in the past 20 years.
Having split my life across multiple cities of residence, I sometimes feel like I’m caught between cultures. People I’ve just met recognize me as a New Yorker from my behavior and mild accent, but I have lived almost none of my adult life in the Empire State, and I rarely visit the city. I’ve spent more of my life in Seattle than in any other city or town, and I’m well-read on the city’s history, yet I’ll never have the swagger of self-styled “natives”, by which I mean local residents, descended from immigrants, who were born and raised here, and who claim superiority over transplants like me. I’ve avoided conflict by avoiding forums where city tenure conveys status, such as local social media groups; besides, many of our elected leaders are transplants, too. Seattle’s current mayor, Katie Wilson, was raised in New York state, the second Seattle mayor from New York in my 20 years here. As one friend from Philadelphia told me when I moved to Seattle in 2006, “nobody here is from here.”
Seattle has been a good place to work, a good place for friends (both preexisting and newfound) and family to gather, and a great jumping-off point for trips by car, train, ship, and airplane. The city has its flaws, but it also has a great community of activists and organizations fixated on solving its most stubborn problems. I’m glad I came here, and while I’m not sure what I’ll do next, what I do next will probably involve continuing to be based here. Here’s to the next 20 years.
]]>I was born and raised in the New York City suburbs, and considering the state of the big city in the 1970s, my parents made the right decision to move out to Long Island. The suburbs had little crime, good schools, and reliable mass transit to take office workers to and from Manhattan. Until I learned how to drive, I also found my hometown to be isolating, with few ways to get around, other than the school bus and car rides to and from my friends’ houses. In 1995, I wrote a letter to Newsday, Long Island’s largest newspaper, in support of a local mother who feared for her child’s safety walking to and from school, on streets with high-speed traffic and no sidewalks; over 30 years later, walking around my parents’ hometown is as hazardous as it was when I was in high school. My current passion for walkable, transit-oriented cities could be traced to my fascination with going into New York City and my desire to get out of my suburban hometown. For years, “no suburbs” has been on my LinkedIn profile, although I still regularly get recruiter messages for suburban-based companies.
I moved to Ballard, in northwestern Seattle, in 2019, and five years later, I bought a car, ending a nearly 17-year period when I didn’t own one. My car-shopping trips took me up to Shoreline, just over the border from Seattle, which has grown so much that it practically blends in with Seattle’s northernmost neighborhoods. From the RapidRide E Line bus that connects downtown Seattle with Shoreline, I can see parks, shopping, new apartments, and education and health centers, built up in a way that should inspire Urbanist Shoreline, a volunteer group whose mailing list I’ve joined. Further to the north of Shoreline is Edmonds, with a charming and walkable downtown, and also a variety of businesses large and small, with a ferry connection out to the Olympic Peninsula. Edmonds is vibrant and flourishing, despite the pleas by some residents not to “Ballardize” it, using the name of my Seattle neighborhood as a byword for the growth and development that they oppose. As the crow flies, Shoreline and Edmonds are as close to me as they are to Seattle’s southern neighborhoods; by transit or by car, they are even quicker to access. Edmonds even has a Winco Foods, which I really appreciate, especially since it has an electric car charger on-site.
The Eastside, unlike Seattle’s northern suburbs, has had very little appeal to me. When I moved to Seattle in 2006, I thought I would be spending some of my time in places like Bellevue and Redmond, but that hasn’t really panned out. I used to watch Steelers games with a group at a sports bar in Bellevue, but because I was renting cars by the hour, I came straight home afterwards, and I later found (and, years later, took over) a group that met in Seattle proper. I thought I’d be going to events at the Microsoft campus, but in nearly 20 years, I’ve been there exactly once, for a university alumni event in January 2007. I interviewed on site at Google’s Kirkland office in 2007; I didn’t get a job there, but had they made an offer, I would have had to confront the choice of having a lengthy commute (without owning a car, at that time) or decamping from Seattle and moving closer to work. I’ve been to Kirkland a few times since then, and it’s grown significantly, but it still has the feel of a small town, by design.
My first couple of light rail trips to the Eastside helped me question my decades-old perception of the suburbs as car-centric and underbuilt. On opening day, I rode the new “Crosslink Connection” out to Redmond, where I had lunch with a few Seattle-based urbanists at a restaurant surrounded by modern apartment buildings, a privately-owned park, and many Microsoft offices. I then continued on to downtown Redmond, where the light rail terminus leads riders into a remarkably walkable and accessible city core. Returning home from the terminus took me nearly two hours, making a Redmond commute impractical, but while I was on the ground, I felt like a commute from downtown Redmond to the Eastside’s many office parks had never been more feasible. The following week, I traveled out to Bellevue Downtown station, and I found that while the station is very convenient to many office complexes, the shopping, dining, and parks of downtown Bellevue are about a mile away from the station, and the pedestrian experience along very wide downtown roads is as bad in 2026 as it was the last time I visited, which was at least seven years ago.
Am I ready to drop my hard prohibition against working in the suburbs? I don’t consider suburban cities to be contiguous with the city just yet, but at the same time, I’m willing to talk about particularly compelling jobs that include occasional travel out to the hinterlands. I also want to make some more trips out to the Eastside, including with my bike, as the weather gets better. There are some amazing trails out there, including some that are right next to light rail stations. Mixing cycling with light rail might make a long commute more bearable. It certainly beats sitting in traffic on a cross-lake bridge, which until recently, was my only practical commute option. Suddenly, I’m more of an optimist about the suburbs.
]]>Like walkie-talkies, MeshCore uses freely available radio waves, although it uses them to send text messages, not audio. I bought a MeshCore starter kit from Amazon that includes a development board about the size of a matchbook, an antenna, a case, and a battery that, strangely, sits outside of the case. (There are other enclosures available, and makers can 3D print their own.) The same board I bought for MeshCore can be reflashed to use Meshtastic, another decentralized network that uses radio waves, but in urban areas of the Seattle metropolitan area, I’ve read that MeshCore is more reliable.
The community around MeshCore is passionate, diverse, and welcoming. My post about MeshCore on Mastodon, a decentralized social network popular with open source developers, went “trending” within an hour, and attracted a reply from one developer plugging his open source project to make MeshCore work with Internet relay chat (IRC) software, which I haven’t tried yet. I’m on two Discord chat servers for local mesh networking communities, Puget Mesh and Cascadia Mesh. The latter group runs the I-5 Project, with the goal of creating a seamess network from Vancouver to Portland. All of this is volunteer-run, and the community is really active in helping new members get set up. As with IRC in the late ’90s, the MeshCore community is small and outside the mainstream, but it has goals that are more aligned with community than with ad targeting.
MeshCore is still an early stage product. Launched in late 2024, its userbase in my area overlaps heavily with ham radio users, many of whom have named their repeaters after their radio callsigns. A battery-powered MeshCore repeater, suitable for outdoor installation, is about $100 as I write this; many users have customized theirs with wind- and rain-proof mounts, solar panels, and larger antennas. No license is required to have a repeater, and they don’t draw very much power or take up much space. My tiny radio came with all the parts in little bags, but no instructions about how to put them together; I found a helpful YouTube video that walked me through the assembly, including some important warnings to stop me from bricking my brand-new radio. I installed the MeshCore app on my phone, and after pairing the radio via Bluetooth, I could see repeaters all around the region and messages coming in on the public channel.
Connectivity on MeshCore, as with any radio-based system, doesn’t work the same way as centralized client-server messaging does. After jumping straight into MeshCore, I saw a flood of public messages and repeater discovery notifications, even while my phone wasn’t connected to the Internet. My house doesn’t have a repeater nearby, so when I’m at home, fewer messages come through. As with FM radios and pagers, if my radio doesn’t have a strong enough signal, I don’t receive messages, and I don’t even know that I’ve missed them. I tried sending my lengthy public key to a fellow MeshCore user on Mastodon, asking them to send me a message, but they told me they weren’t able to get through. The primary public channel can be noisy, with chatter centering on new users testing the system and occasionally talking about the weather. People can also set up public “hashtag” channels, using the same #channel naming scheme that I remember from my IRC days. There are also ways to set up private channels, hidden behind a password, and encrypted for security.
I can imagine MeshCore being useful if a disaster knocks out cell towers, although I’m not sure that people’s home-mounted repeaters would be more durable than professionally built 5G towers. One other key use case would be protests, where carrying around one’s mobile phone could lead to their identity being compromised. There are several devices, such as the Lilygo T-Deck Plus, that include the LoRa long range communication equipment in a case with a hardware keyboard and a screen; the T-Deck Plus reminds me of my old BlackBerry, which had its own durable, pseudonymous messaging service. Carrying a T-Deck could become as useful for operational security as burner phones have been.
Is MeshCore fun? Absolutely. Is it going to displace mass-market, commercial messenger software like WhatsApp, or open source messengers like Signal? Almost certainly not, considering that you need to buy special hardware to use it. The fact that I can use this service for free, with a passionate and supportive community stretching from the city to the hinterlands, means that I might have another outlet if I ever need to communicate in an emergency without mobile data being available. Even if I never need it, I still think I’ve gotten a good return on the $30 I’ve spent on joining the mesh.
]]>Since late 2019, I’ve been visualizing elections in King County, Washington, using data that shows votes in each of thousands of the county’s precincts, including over 1,000 in Seattle. It started as a way for me to improve my Tableau calculation and mapping skills, and it’s connected me with campaigns and journalists whose work I can support directly. At this point, I typically reuse the same Tableau workbook from one election to the next, swapping in updated precinct shapefiles and election results, which the county provides in standard formats. The most time-consuming process is going through every race, of which there can be hundreds in an election, to ensure that candidates have contrasting, appropriate colors and that all the names are encoded correctly. Tableau Public kindly hosts all my visualizations for free, but as Tableau has been absorbed into Salesforce and as its strategy and personnel have changed significantly, I’ve wanted to build visualizations in other ways. During my most recent job working on open source Jupyter projects, I built a Jupyter notebook to convert shapefiles and results into interactive visualizations, but they didn’t look as good as Tableau’s output did, and the resulting files were so large that they could have become a financial burden for me to host. I decided to see if Claude Code could build some maps for me.
I responded to my friend’s challenge by plunking down $20, plus tax, for one month of Claude Pro, and installing Claude Code on my aging MacBook Air. My first attempt to create a visualization led Claude to build an app that tried to download files directly from county web sites, which didn’t display any usable map, and which wasn’t fixed with any amount of follow-up prompting. I instead gave Claude Code access to copies of the county’s shapefiles and results files that were already on my hard drive, and I asked it to produce static output that I could host using an Amazon S3 bucket and Amazon CloudFront as a distribution network. It asked me a few follow-up questions, like a good job interview candidate, before displaying a series of progress messages and animations. It explained what it was doing as it took over 7 minutes to produce code for me. Less than 30 minutes after I installed Claude Code, I had a web page that displayed election results in a color-coded, readable, mobile-friendly way. This was great progress, so I decided to think bigger, and this required a lot of follow-up work with Claude Code.
I’ve heard a lot about AI coding assistants as a disruptive technology, particularly for companies that are used to hiring junior developers. In my experience, Claude Code behaved like a friendly, obedient junior developer with a very restrictive work schedule. With the Claude Pro plan, I have session usage quotas that reset every 5 hours or so, and weekly usage quotas as well. After just four sessions, I had already used a quarter of my weekly quota, and one of those sessions ended abruptly, because I had reached my usage limit. Although Claude Code lets me queue up multiple prompts, it’s not really conducive to the “flow” state that’s essential to productivity. Even simple actions, like find-and-replace operations that modern text editors can do near-instantaneously, took 10 seconds or longer in Claude Code, distracting me from more complicated tasks. Especially when we were getting started, Claude asked me a lot of follow-up questions out of an abundance of caution; unlike a junior developer, it didn’t seem to learn from my responses. I made a few explicit suggestions, such as to use U.S. English spelling, that were saved for later. I was also given the option to be asked to approve every file change and script run that Claude Code chose to make, or to give AI free rein to do whatever it deemed necessary, with seemingly total access to all my personal files. Unlike a human developer, Claude has no accountability for its actions; had it done something destructive, I’d have to rely on my hard drive backups to bail me out. If I needed to rely on Claude Code as a full-time employee, I’d almost certainly need to upgrade to Anthropic’s “Max” plans, which start at $100, or I’d need to pay for even more usage on top of that. As with all hosted software providers, Anthropic could lower my usage limits, suspend my account, or raise their prices at any time; in other words, my vibe coding assistant could demand — and receive — a raise at will.
Claude Code works fast, but it doesn’t think big, and it often makes errors, especially as a front-end developer with no eyes and no fingers. Particularly on mobile web browsers, my election visuals sometimes had elements that were too small to see or touch, that appeared overlapping each other, or that had nonsensical yet correctly spelled labels. It chose green and red for ballot questions that were answered with “Yes” or “No”, unaware that this is a bad visualization practice, making the map unreadable for people with red-green colorblindness, which includes up to 8% of men and 0.5% of women. (To its credit, Claude Code included some accessibility markup in its output.) Its HTML output included scripts and styles inline, without using semantic markup or class names much. This sets a bad example: Claude Code can edit its own code just fine, but when I looked at it, I didn’t like what I saw, and Claude Code only improved its style when I explicitly told it to. Sometimes, when I asked it to solve one problem, then a second problem, the first problem reappeared. As part of this experiment, I delegated to Claude Code entirely, but in a production environment, an operator really needs to blend AI-driven coding with some well-informed guidance and knowledge of best practices.
I’ve used a few different AI tools for coding assistance, including developer extensions like Jupyter AI, which my team at AWS built and still maintains. They can produce code so quickly, they impress and sometimes scare me. They produce output that superficially does the job requested, but that falls apart in the face of scaling questions, human usability, and security challenges. They do not provide any accountability or enough insight into new and complex problems; they can produce output that doesn’t answer the question they were asked. They make fantastic prototyping and experimentation tools, and as an experienced developer, I think they do a great job of automating the more boring and repetitive tasks that other people have already solved. I also appreciate that once my one-month Claude Pro subscription ends, I can keep using Claude Code with Anthropic’s models hosted by Amazon Bedrock, which charges me a metered rate based on usage, without the same quotas as Anthropic’s own Claude Pro plan. Perhaps, when computer prices normalize again, I could host my own AI models locally. The future’s looking bright and productive, whether or not I have a job in it.
At publication time, I owned shares of Amazon and Salesforce; the latter company owns Tableau. None of this article about using large language models (LLMs) was created or edited using an LLM. All em dashes were authored by hand.
]]>Surveillance tech, when used by trusted, well-regulated organizations in responsible ways, can indeed save lives, as recruiters for these companies like to say. It would be great if police officers wore body cameras, so that abusive behavior gets documented and corrected. First responders would benefit from having high-quality camera drones to get better views of dangerous situations. Unfortunately, we are not in a moment where law enforcement agencies, particularly at the federal level, are reliably acting in responsible or accountable ways; to the contrary, leaders have assured rank-and-file members that they can do whatever they feel is necessary, without consequences to themselves. The trust I had put in these agencies, that in some cases was being rebuilt over decades, is at a new low right now, as law-abiding observers are arrested, attacked, harassed, and otherwise victimized by officers that I’d been told are ensuring “public safety.” Working on surveillance technology, right now, exposes members of marginalized groups in the U.S. to even more danger than they’re already in, whether they’re legally entitled to be in the country or not.
Optimism about tech’s impact on the world doesn’t always match reality. Fifteen years ago, during the Arab Spring, revolutionary groups used social media to coordinate uprisings against authoritarian governments that were indifferent to everyday citizens’ needs. While many changes in policies and leadership resulted from protests in that year, in many cases, countries with popular uprisings in 2011 are still in bad shape in 2026. It’s been shown repeatedly that digital media are more effective at stifling dissent than they are at enabling it. When I started using the Internet 35 years ago, basic services like e-mail and the web were highly federated, but today, even the most tech-savvy Americans typically rely on the same small set of giant companies for all of their communication needs, and bristle at the idea of, for example, moving their private messages away from services that feed into ad tech algorithms. The same is true in most countries. Homogeneity in tech makes it much easier for people to communicate among themselves. When tech executives so readily capitulate to authoritarian demands, surveillance becomes a standard feature of key services provided by the private sector.
I speak from a position of privilege when I exercise my choice to opt out of yet another industry that employs thousands of talented techies, and whose companies fund useful, non-creepy projects with some of the revenue they make from surveillance tech. I don’t believe that an employee at a surveillance tech firm can effect major positive change from the inside; that’s optimistic, but ultimately, conscientious employees will reach a point where they have to quit or choose to disregard the effects their company has on the world, so that they continue to get an income to support their own personal needs. I also admit to hypocrisy about these companies; I’ve worked on ad tech and on tools that can be used in surveillance tech, including reportedly by Immigration and Customs Enforcement (ICE), and I have indirect investments in surveillance tech companies, through mutual funds in which I own shares.
I believe that much of the tech industry’s capitulation to the current presidency is out of fear, rather than out of enthusiasm. Helping to fund a new White House ballroom, supporting the first lady’s documentary about herself, and attending inaugural balls are all ways to try to stay in our authoritarian president’s good graces, at least in theory. Even in this supposed age of government efficiency, government contracts are critical to many companies’ revenue and long-term success.
For a deeper analysis about surveillance tech, with a broader definition than what I’ve used in this post, I recommend Shoshana Zuboff’s very good book The Age of Surveillance Capitalism. Written in 2019, it was an accurate look at the inequities of the tech world at that point in time. I hope that Prof. Zuboff, or someone she endorses, is working on a follow-up book covering the troubling governmental abuses that I expect to continue for a few more years, at least.
]]>Did anything of value just get accomplished?
I’ve spent nearly a year not working in tech, observing how the current AI mania has affected businesses of all sorts. Leaders are full-throated in their pursuit of “efficiency,” whatever it means to them. As I started my tech career in 2003, the pursuit of efficiency led companies to try outsourcing many processes, including software development, to employees or contractors in lower-cost locations. From my observations, these have been a mixed blessing for companies: there are plenty of opportunities for fresh university graduates in, say, India, but the most talented people want to transfer to places like the United States, where pay and opportunities for advancement are much better. Now, the trend is to try cutting humans out of companies altogether. Is it working?
As I unwrap the AI-drawn trinkets from an advent calendar I was given, I smile, knowing that a well-meaning friend wasn’t deterred by the obviously AI-generated artwork on the packaging. The first and most obvious targets of the current AI boom are the lowly paid, often offshore, artists who design the assets to promote products — what’s often called “creative”, as a collective noun. Cheap creative has been a hallmark of cheap products and knockoffs for generations; now, someone can run a few prompts and reduce their costs to a tiny fraction of what human workers used to charge. Low-quality “chum” banner ads, emotionless illustrations, product apps and websites done on a minimal budget — all of this creative work can now be done with AI, and although customers might notice, the marketplace of low-budget products welcomes it in.
“Content”, as some companies broadly call their written prose, can be generated easily using large language models (LLMs). Human-written articles on content farms are now fed into LLMs’ training data sets, and web searches for topics like household tasks and recipes now return both content farm articles and the AI-written articles that are derived from them. Duolingo, a language learning app that I’ve used for over a decade, is public about using LLMs to create lessons more quickly. The quality and cultural relevance of Duolingo’s lessons continue to be poor — they’ve taught me to say “the cat eats seven strawberries” in many languages, for example — but if the key performance indicator is quantity, then LLMs can help the company hit its metrics. For an extra fee, Duolingo Max will even let me have a near-real-time, LLM-aided, conversation with one of its characters in a language of my choice. It’s somewhat useful, but mostly creepy.
It’s fun to laugh at and dismiss the confident-sounding nonsense text that LLMs spit out; one of my old colleagues, Vincent, dedicates much of his Bluesky account to this. I’m concerned that I also see AI-generated creative being produced by local volunteer organizations and even government agencies. The obvious marks of non-human art and writing damage the credibility of anyone who puts a real brand on synthesized artwork. I’ve also talked to recruiters and job-seeking developers who believe in the pervasive myth that the “ATS”, an AI-powered talent scoring system supposedly used by many companies in the same way, assigns a single numeric score to every applicant’s resume, so a resume should be rewritten primarily for the ATS, not for humans, to read. Some companies now do initial interviews, particularly for entry-level jobs, using fully automated recruiters. If I were invited to such an interview, and I couldn’t spin up an AI bot to speak to a recruiter bot on my behalf, I’d walk away.
On the other hand, well-implemented AI has been incredibly useful since long before ChatGPT went viral in late 2022. Travelers like myself would be lost without machine translation; for example, on my most recent visit to Japan, several restaurants had smartphone-based ordering interfaces, and they relied on translation features built into my phone’s browser, which in turn have used AI models for many years. We’ve all come to accept automated speech recognition and synthesized voices in the products we use; for short text phrases, they listen and speak well, they respond quickly, and they’re far cheaper than hiring professional operators and voice actors. I’ve even seen people try to use their phone or earbuds as a real-time translation tool, which doesn’t work particularly well right now, but I’m optimistic that these features will improve over time. Not too long ago, YouTube’s automated captions were hilariously inaccurate; now, they are incredibly good at compensating for low audio quality, background noise, thick accents, and the use of proper names. AI models make those captions, very useful for accessibility and search indexing of videos, possible.
I also think a lot about the “creepy line,” as coined by Eric Schmidt, then CEO of Google, the point at which technology becomes too off-putting to keep using. When I was attending public weekly meetings for Jupyter, the open source project I worked on from 2021 to 2024, one person had an AI note-taking bot that showed up as a second attendee. When neither the attendee nor the bot responded to questions about what the bot was doing, the community collectively decided to ban third-party recording bots from listening in on meetings, including discussions we considered “off the record” and excluded from the published recordings. Earlier this year, I considered interviewing with a company that makes a bot that listens in on meetings and writes incredibly detailed summaries not just of the text content of the meeting, but also of who spoke for how long, and how attentive people on camera seemed to be. The bot lets individual people opt out of its reports, and makes its own presence known in meetings, but despite these good practices, I felt like such a bot was over the creepy line for me. When I’m given the choice to opt out from recording, I take it, so I felt I wouldn’t be an effective developer of something I wouldn’t want to use myself.
I consider individuals’ AI voice recorders and “smart glasses”, with built-in cameras and microphones, to be over the creepy line. At some point in the near future, I see myself in a situation where I’ll ask someone to remove their prescription spectacles because I don’t want them to be recording me. (Many frames have lights to indicate when they’re recording, but some quick electrical work or a piece of tape will render them useless.) If they refuse, I’ll leave the conversation. People have called me old-fashioned for exercising my right not to consent to the recording of private conversations, but I’m sticking to it. Facial recognition is so far over the creepy line that even Facebook shut its implementation down in 2021, although unsurprisingly, they resumed using facial recognition last year. Companies and governments still use it; I opt out of the facial recognition camera at the TSA checkpoint when I take a domestic flight, but when I’m at a private business, they generally have the right to check that I don’t look like someone they’ve previously banned. There’s no way to know how many Ring doorbells in my neighborhood clock my movements as I walk past a stranger’s front door. What’s creepy to me might be useful to someone else; for example, many of my former coworkers learned English as a second language, or had cognitive issues that made it harder to process speech in real-time, so AI tools would help them understand speech more quickly, even in a private chat. If someone really has a compelling or compassionate reason to use AI, I’d reconsider where the creepy line lies. AI devices that use locally-hosted models, without sending recordings to be archived forever in a corporate data center, might land on the good side of the creepy line.
Is AI a bubble? Yes, in a similar way as the web was a bubble 25 years ago. The dot-com bubble burst, a lot of investors lost a lot of money, a lot of talented people lost their jobs, and web technologies are more dominant and more lucrative now than they’ve ever been. AI technology predates ChatGPT, and it will continue to improve with time, no matter what happens with investors and data centers. Twenty-five years from now, there will still be AI and LLMs, and technologies derived from both, and there will be ever more uses for them, which will require talented and, perhaps, ethical people in charge. Maybe we’ll even have government regulations to keep them in line. A lot can change in the years ahead.
Author’s note: Although the text above includes em dashes, it was all human-written; no LLM was used in the production of this article. At publication time, I owned shares of Amazon, which owns Ring.
]]>One of the big motivators for my break is that I had had an unhealthy relationship with work. As far back as college, when something negative or personally difficult came up, rather than dealing with the issue at its core, I tried to power through it with work. On the surface, that approach seemed to be good for me: I graduated from college a little ahead of schedule, and I’m proud to have worked on challenging projects at several different companies. As good as this work was for shareholder value, though, it wasn’t always good for my mental health, and until recently, I had rarely considered that a concern.
It’s very likely that 2025 will be a year where I’ll have produced nothing of commercial value; that’s fine with me. Not working at a company means that I’ve been working on myself instead. The time I invest pays dividends in different ways than salary, bonus, or shares of stock. I’ve challenged myself to exercise more intensely this year than I had in the past, resulting in better fitness, and keeping my weight under control at a time when I’m not riding my bike to an office every day. Instead of refining my programming language skills, I’ve been honing my interpersonal skills, learning to be a better and more supportive companion to people around me personally, not just those who are going to write an annual review for me. Whether it’s physical exercise, language skills, social skills, or my chess game, regular practice and helpful feedback make me better. When I’m ready to take another job, I will have a better sense of boundaries between my work life and my personal life.
I had two moments last month, in the same week, that reminded me of how removed I am from the working world. First, I donated money to a public radio station, and they very nicely invited me to take a tour of their office to express their gratitude. As I told my guide, the brief tour was the most amount of time I’d spent in an office building this year. I felt very confused and out of place, like I didn’t even belong in an office environment after less than a year away from it. Secondly, I took a call from a corporate recruiter for the first time in years. The job seems interesting, although again, I found myself very unfamiliar with this type of call. It’s similar to a situation I’ve found myself in on some of my travels this year, where I have to jostle some stale spoken language I’d studied out of the attic of my brain, so that I can understand what’s being said to me in a foreign country.
Also recently, I reopened Katrina McGhee’s book Taking a Career Break for Dummies, which I read last year, to review the section towards the back about ending a break. McGhee took 20 months off work before coming back into the workforce as a self-employed career break coach. She advises people to take one-sixth of their break time to date as a time to focus on re-entering the workforce; for example, if someone’s been away for a year, the re-entry process might take two months. A big part of that process is introspection, identifying things that are of great importance to the person on a break. Most of my introspection is going to remain private, but for a future job in software — and I don’t foresee myself changing to a new career in the near-term — I came up with these priorities, in no particular order:
I have some more travel coming up, but as I write this, there’s nothing on my calendar for 2026: no trips, no work, no classes, nothing big that creates an obligation. I’m having a great time working on myself and rebuilding what has broken down over the past 22 years. Nothing forces me back to the workforce, but I’m curious about my options.
]]>I studied Japanese back in college, and in 2002, I spent a summer studying abroad in Tokyo. I stayed with a host family that had dial-up Internet, which cost a fair amount of money per minute, and I had no mobile phone of any kind, so I had to be self-reliant when I was on the go. The fanciest gadget I carried was my Casio EX-word electronic dictionary, which came with a stylus and sort of recognized my handwritten characters; at about $100, it was such a great tool that I brought one of my classmates to the same back-alley shop in Akihabara so he could get one of his own. I’ve been back to Japan three times since as a tourist, next-most recently in 2018, and on each occasion I’ve had a smartphone with apps for translation, mapping, and web browsing, that make me feel like I’m cheating, compared to my 2002 experience.
This time around, I came to Osaka, a city I’d never spent any time in. I had tickets for Pokémon Go Fest and Expo 2025, which were being held nearby, and my old iPhone 11, which I use as a travel phone. I bought an Airalo eSIM with 5 GB of data, which turned out to be just enough for my lengthy Pokémon Go sessions, with enough data left over for light web browsing and ChatGPT usage.
Having a background in Japanese certainly helped me a lot; it would take lots of time and energy for ChatGPT to teach the fundamentals of etiquette, food, payment, and non-verbal communication skills that make interactions go more smoothly. I had also spent some months brushing up on my Japanese using Duolingo, which, like most translation apps, does an OK job of teaching me words and sentences, but offers no cultural education at all. ChatGPT fills in a lot of cultural gaps. For example, while Duolingo’s rigid lesson plan was teaching me to talk about kings, queens, and knights for some reason, I could ask ChatGPT to walk me through the domain-specific vocabulary I’d need to, say, buy commemorative silver coins at a dealer. The AI app can also role-play conversations, something that I found valuable when I was taking Japanese classes in college, and that Duolingo only offers as part of its super-premium “Max” subscription that costs nearly as much as ChatGPT Plus does. ChatGPT even could teach me a few phrases in Osaka’s distinctive Kansai dialect, something that I barely remember from my college class days, and that Duolingo doesn’t teach.
I can tell that ChatGPT’s data set includes a lot of the expatriate forums that I found useful on prior trips to Japan. It can tell me about popular local brands similar to over-the-counter medicines that I use at home, so in case I run out on my trip, I know what to look for at a pharmacy. (Google Images can help me find what these brands’ packages look like, helping me even further.) It can recommend stores to buy particular goods, and can give me suggestions about how to ask a clerk for a particular item. It can also give answers that are respectful of local customs. One of my old coworkers went to Kyoto some years back, for example, and despite having Google Translate, was frustrated when they couldn’t figure out how to get “free refills” at a tea shop there, as they were used to at home. ChatGPT tells me that, in Japan, traditional tea shops charge per pot or per cup, with free refills generally not offered. Conversations go more smoothly if I better understand whether I’m about to ask the right person for something I can reasonably expect to receive.
ChatGPT’s photo recognition was a real gem during this trip. Google Translate supports translating text inside an image, which is useful at certain museums, where professionally printed signs describe, in brief, what I’m looking at. Shrines, temples, and historical sites include some plaques with essay-length texts written on them; in that case, I don’t want a literal and complete translation on my phone. Photographing a text-dense plaque at Shitennoji temple, for example, led to ChatGPT explaining that I was looking at a historic example of drainage pipework, without having to read the complete history to understand this. At the same temple, photographing certain objects outside produced explanatory responses that concisely described their purpose and significance. Unfortunately, inside some buildings, photography is not allowed, and I didn’t test whether I could get away with taking pictures by saying that they were for my virtual assistant, rather than for my camera roll. Photo recognition responses from ChatGPT, like many AI tools, are confident, but not always correct; when I sent ChatGPT a photo of the Hanshin Tigers’ minor league stadium, for example, I was told that I was at the team’s major-league facility, Koshien Stadium, which the minor league field is designed to resemble.
Where ChatGPT really shone on this trip was in recognizing handwritten and very stylized Japanese text, including elaborate signage at restaurants and bars. One bar in a popular nightlife area had a lengthy handwritten sign outside, with the only English text reading “For overseas people” and “Only the Japanese menu”. In seconds, ChatGPT clarified that at this bar, patrons are not allowed to use mobile phone apps to translate the menu or communicate with staff — something that, if it were posted outside in English, would probably incite a lot of harsh reactions online. I’ve been at restaurants in many countries, including the U.S., where menus use so many esoteric words or design elements that even a fluent speaker of the local language might get confused. An AI assistant won’t be as well-informed as a waiter will be, but it can probably answer a question more quickly, without awkwardness or judgment.
My next international trip will take me to Norway and Sweden, where I’ve spent effectively no time before, and where English proficiency in cities is excellent. Nevertheless, I’ll still use ChatGPT to improve my local knowledge, and to feel more confident in my travels. Especially when the barrier to entry is so low — just the cost of a local eSIM, which could be as cheap as $5 — it’s worth a try.
]]>During the first half of this year, I’ve had a mixture of warm-weather trips, nostalgia trips, and trips to new places. During the winter, I went to Kauai for the first time in over a decade, to San Diego for the first time since 1986, to Palm Springs for the first time, to Pittsburgh for the first time since COVID, to Vancouver Island, and to Osaka, my first Japan trip since 2018. By the end of June, I will have been to Virginia, to meet a couple of friends who I’ve known since our IRC days in the late 1990s but who I’ve never met in person, and to Washington’s Methow Valley. More trips are coming up in the second half of this year, too. During these trips, I’ve been doing a mixture of exploration, meeting up with people I haven’t seen in anywhere from a few months to 20-plus years, and introspection.
Despite all this movement, I remain planted in Seattle, my home of just over 19 years — a longer span than anywhere else, including my first 18-plus years when I grew up in New York. I keep planning trips as one- to two-week vacations, with the strong intention to come back to Seattle when I’m done. I have a home in a walkable, bikeable, livable neighborhood. I have friends and family who live nearby. I’ve been able to dedicate more time to local volunteer causes, without business meetings or conflicts of interest to get in the way of housing, transit, and walking advocacy. The closest thing I’ve had to a full workday involved a 6:30 AM shift with a local volunteer group, handing out coffee, doughnuts, and swag to bike commuters, cycling 9 miles to an afternoon baseball game, then cycling to the group’s afternoon happy hour. I’ve also been able to travel with a few volunteers to Pike Place Market to advocate for the market to be more pedestrian-friendly, which has since been tested with great positive impact for visitors on foot.
As the tech industry continues to lay off workers, leans into artificial intelligence as a substitute or augmentation for human work, and deals with political challenges, I haven’t spent much of my first 6 months doing technical work. I typically travel with my phone and an iPad. The closest I’ve come to software development work is some light revision to a few silly Discord bots that I’ve deployed on a small social server. I’ve been reading books like Abundance and The Tyranny of Merit to try to make sense of the present and future of U.S. politics. I also picked up How to Retire, even though I still reject the “R-word” myself, to better understand how professionals see the financial and emotional challenges of life without work. Aside from Pokémon Go and slow-paced online chess games, I haven’t been playing many video games during this break. I still get recruiter mail from time to time, and I’m open to picking up a new job, if something really compelling comes up.
I’m still staying physically active with a mixture of walking, biking, and one-hour workout classes at a local gym. I’m keeping my brain active with language learning, word games, and regular meetups with friends and fellow volunteers. I’m also continuing to talk with a therapist on a regular basis, which helps me deal with the vastly increased time that I’ve spent thinking about my feelings and emotions during this break. I’m very fortunate to be able to spend this time healing and recovering from both the accumulated stress of a two-decade software career and the events of this decade to date. Other people who have taken breaks have been very positive and open about their experiences; I’d be happy to talk with anyone considering a break of their own.
]]>