I pitched this idea to Antler in NYC and was accepted into their Fall 2025 cohort to validate the thesis.

Over the course of building and testing Trailway, I learned four hard truths:

1. The pain is real but not urgent:
   • PMs feel the pain, but it’s not urgent. Most saw hiring more PMs or adding QA as a future solution.
   • Engineers feel almost no pain because PMs or QA handle it.
   • QA teams see new feature testing as baseline work, not a top priority.
2. There was no clear owner:
   • In PM/Eng-only teams, ownership of pre-release testing is ambiguous. Sometimes engineers thoroughly test new features, making it easy for the PM to approve the release. More often PMs delegate testing to Engineering and when that fails, they take on testing by default.
   • When QA teams exist, they own execution. But depending on their experience with the platform, PMs shadow-own by double-checking the work.
3. No clear budget or power to bring on a new tool:
   • QA lacks budget and influence to bring in new tools.
   • PMs hesitate to formalize responsibility by bringing in new QA tools.
   • Engineers have higher priority work to do.
4. The solution is messy and context-dependent:
   • Every company’s SDLC process was different, making a one-size-fits-all solution difficult.
   • General prototypes weren’t attracting early adopters.

In summary, new feature release testing stayed broken because the pain was uneven and non-urgent, ownership was ambiguous, decision-makers lacked budget or incentive to adopt tools, and the underlying workflows varied too much for a simple, compelling solution.

Without a co-founder to share the load and challenge these learnings, the constant friction slowly wore down my conviction and made it hard to keep going.

As I learned more about the complexity of the QA market, I chose not to pivot, since it pulled me further away from my original goal of helping product managers. For more insight into the QA market landscape, read my post: The Quiet Crisis in QA.

I’m grateful for the chance to explore a problem I care about. Thank you to everyone in the Antler program, the people who took meetings with me for research and advising, and my friends and family who were my sounding board throughout.

This post is meant as a high-level overview. For deeper questions, feel free to contact me—details are on the home page.

What’s surprising is there aren’t many. Even A16Z’s “The Trillion Dollar AI Software Development Stack” barely touches on the advances in QA and lists no companies in that space. This gap isn’t because people aren’t trying, it’s because they are finding out how hard automating QA is.

The discovery

I realized this after starting the Antler accelerator. I joined building Trailway.ai, an AI agent that runs through your feature before release, finding design problems and bugs. While outlining my QA approach from past roles, I found it easy to write down obvious things like broken links and buttons. It was a lot harder to define what “quality” and “good” looked like. There are many ways issues appear, and all I could say was “I know it when I see it”.

Antler founders were excited by the idea of AI finding their bugs. Yet, they too struggled to say what they needed. Speaking with more people at different sized companies, I realized they too had a problem but had trouble articulating their needs. It was the early sign of a tarpit idea.¹ So, I went to work better defining the problem to solve to help find a solution.

How QA Plays Out in the Real World

This is a simplified view, but it’s helpful to start here. QA exists to make sure your app works the way it’s supposed to. Bugs are what happens when it doesn’t. So the job of QA is to find bugs before your users do.

Bugs aren’t just when something doesn’t work. It could work, but not look the way you meant. Sometimes it’s worse. It works the way you meant, but breaks something else in the process.

The most subtle bugs aren’t technical at all, they come from people misunderstanding each other. A feature works as built, but it doesn’t match what the people defining the experience actually wanted.

When you’re solo, most bugs are caught quickly because you test as you build. Add another person and misunderstandings start creeping in. What to build gets misread, conversations get forgotten, and assumptions drive decisions. The risk shifts from how you build to how you communicate. As teams grow, these challenges multiply. Bug-finding becomes even more about communication and coordination.

If all of this sounds complicated, that’s because it is. But that’s not what people want to hear when they are trying to solve their QA problem.

QA is hard

Folks are now used to Vibe-coding platforms like Loveable, that make it simple to build and deploy a web app quickly. Just say what you want and watch AI code your site for you. This expectation for simplicity shows up in QA as well. Let’s call it Vibe-QA’ing.

But as most people learn after using vibe-coding tools, software development is hard work and these tools don’t solve that. They’re great for prototypes or simple sites, but struggle as your app inevitably grows.

It’s safe to assume Vibe-QA products will face similar struggles, and maybe more due to the nebulous nature of QA. Even if these products work on small-scale examples, it begs the question: Who is their ideal customer?

QA is often an afterthought

Small teams don’t focus on QA. Their aim is growing the company by solving more and more problems for customers, not delivering bug-free software. When needed, they focus on testing critical user paths only, catching bugs quickly, and skipping automation.

Larger teams have bigger stakes. Each engineer only knows a small slice of the product, so bugs spread wider and get more expensive. With paying customers, contractual agreements, and brand reputation on the line, structured QA and automation become the only way to keep quality under control.

It’s often why companies finally decide to prioritize QA. It could be that their business stakeholders are tired of manual testing all the time. Maybe their difficult release process burns out the engineers. Or, it could be a major issue that causes revenue impact causing testing to be a top-down mandate. Whatever the reason, now the company has an uphill battle to “do QA”. Fortunately for them, this has been the pattern for a while, so there are a lot of tools that help.

Big players solve most things

To no surprise, companies have been working on solutions to improve testing for 10+ years. Some have close to half a billion in revenue. These tools like SmartBear, Tricentis, and Browserstack are comprehensive testing suites.

That doesn’t mean they work perfectly. It’s just, there is limited room for re-inventing software testing in a breakthrough way.

Most smaller players are re-packaging what the big players already solve

Every smaller player I’ve seen is tackling a different piece of the problem. But, almost every one of those pieces is already owned by the big players.

Some go after test case management. They help teams organize, reuse, and maintain their tests. These tools are about keeping what already works from breaking. It’s a stable market with old tools like TestRails dominitating and newer upstarts like Qase.io finding room around the edges.

BrowserStack Test Case Management Marketing

Others focus on visual testing. They catch unexpected visual changes before they slip through to users. This is done by comparing screens between versions to see if anything looks different. It’s easy to use and surprisingly effective. Meticulous.ai and Chromatic.dev are the main ones here.

BrowserStack Visual Testing Marketing Nov 2025

Then there’s bug reporting, for manual testers. These tools make it easier for people to share what’s wrong with all the helpful information an engineer needs to fix it. jam.dev and marker.io are the main ones here.

Jam.dev Marketing

The biggest wave, though, is record-and-playback automation. These tools provide a visual builder that watches your actions in the browser, turns them into a repeatable flow, and lets you add simple checks along the way so the system knows whether each step is working. The result is a set of flow-based tests that sit on top of code-based tests but are far easier to create and maintain. They help you catch regressions in areas of the app that used to work. With LLMs involved, the automation gets even smarter, figuring out what to click even when the UI changes slightly. Startups like RainforestQA and QA Wolf are going after this, along with a wave of YC companies like Momentic, Spur, Stably, Docket, Qualgent, and others.

RainforestQA Marketing

Most of these companies have “AI” mentioned big and bold on their landing pages. But so do the big players.

The new wave of QA companies

A few places are working on novel approaches to QA. They aim to test the whole app with minimal work. They fall in the gray zone of “Vibe-QA”.

One approach is using browser agents. These tools are using exciting new breakthroughs in AI models that can use your browser, like a real person.² They explore your web app trying to break things. One of the few places fully focused on this is Propolis. It’s like the “monte carlo simulation” of testing, hoping one of their agents will find an issue.

Propolis Demo Video

Another approach is session replay. These tools capture real users’ sessions directly from your live site and automatically replay those sessions on both the old and new version of your app to spot any new errors.³ Companies like meticulous.ai and older open source tools like GoReplay are approaching testing this way.

It’s exciting to see new approaches appear. We’ll have to see if they fundamentally change QA. But, either way, it’s clear that choosing and using these tools is complex and becoming its own problem.

It’s simpler to use Q-as-Service (QAaaS)

With all these tools and options, someone has to know how to pick and use the right ones. Usually, that means hiring a QA team. But finding people with the right expertise and then managing them isn’t easy.

So instead of building that expertise in-house, many teams outsource it to specialists. Take RainforestQA, for example. They sell a product, but many of their customer success stories mention the RainforestQA team doing much of the heavy lifting.⁴

QAWolf took a similar route: they started as an open-sourcing testing framework then pivoted to a closed-source services company. Now, you pay them for services, while their in-house QA team uses their platform to do said services.⁵

These companies end up feeling Palantir-esque: they build complex but powerful software, then wrap it in consulting and sell the full package.⁶ It’s not necessarily a bad thing. These teams can move faster and know the domain inside-out. But it does tilt the balance of power. The closer these companies get to a customer’s workflow, the harder it becomes for that customer to leave for better alternatives.

What’s exciting

What excites me the most isn’t breakthrough new technologies that re-invent testing. It’s creating exceptional user experience with the solutions we’ve already seen.

A great example of this is Jam.dev, a bug reporting tool that came seemingly out of nowhere. To their credit, they have hustled through thirteen launches and pivots over five years to achieve success. They don’t even tackle “QA” directly; instead, they focus on making bug reporting effortless for anyone on a team. In doing so, they’ve managed to unbundle a small slice of what the big players offer, turn it into a delightful user experience, and expand their audience beyond engineers to include roles like customer success.

It’s a reminder there’s still room for creativity. You don’t need transformative technology, nor rebuilding the whole QA stack to create new value. Sometimes, the win is making one overlooked piece of what exists truly great.

What the future holds

The established players in the QA space will keep doing what they do best: adding AI features that make their tools incrementally better. Think auto-generated test case descriptions, self-healing tests that adapt when there are small ui changes, and so on. They’re focused on making QA teams faster at repetitive stuff, rather than groundbreaking technological changes.

For new companies trying to break in, it continues to be an uphill battle. They’re not just competing against big players. They’re also competing against each other in an increasingly crowded space. Incumbents have a real advantage here: they bundle everything into one product, and since QA isn’t a core differentiator for most businesses, companies usually just want whatever’s easier in the long run. That typically means sticking with the established player.

On top of that, most of these new entrants offer nothing truly differentiated or hard to copy. So, capturing significant market share is really hard.

Some of these new players promise that AI will test everything, which sounds great. But, there’s a ceiling they’ll likely hit around intent. That is, knowing what the product should do, edge cases that matter, user expectations, and more. AI can scrape clues from scattered documents, chats, design files, notes. But it can’t pull insights from hallway conversations, calls, past decisions, or the knowledge people never bothers to put down. These tools might get good at flagging potential problems, but they won’t replace human judgment.

This creates an interesting problem for these tools: they end up being additive to existing workflows rather than replacing them. They add another thing to check, another tool to integrate, rather than making what people already do dramatically faster. And additive is a much harder sell than “this replaces hours of manual work.”

That said, we’ll see real productivity gains. Testing cycles will speed up as agents propose fixes for issues found. The dumb stuff, like broken buttons, will get caught earlier saving everyone time. Bug triaging will move faster because the right debugging data will be captured automatically. And QA teams will be able to push coverage much further by using record-and-playback tools and coding agents to generate test code.

So where does this leave us? I think the perception that QA is falling behind as teams ship faster is a bit misleading. QA tools are getting better too. It’s just less explosive growth. We’re seeing incremental progress rather than a revolutionary breakthrough. Maybe we haven’t yet seen the breakthrough that fundamentally changes QA. It’s just not the kind of thing that makes for great headlines.

A marketplace for booking NYC bars and restaurants for group events, starting with large-party events like birthdays. Think “Opentable for large groups” where you select the packages and sign the contract up-front, rather than via forms and emails.

How did we form the idea?

While having burgers at Jackson Hole in the Upper East Side, my cofounder Jin and I imagined renting the restaurant to throw our own party. But there’s no easy way to book it. That gap reminded us how painful it was to book wedding welcome events. Unclear pricing, unclear venue availability, and slow replies. We saw the same problems when we and our friends tried to book NYC bars and private dining rooms for birthday parties. We realized we could make bookings easier for birthday parties, weddings, and other group events.

What did we try?

We marked the start of our company by creating a Slack workspace on June 12th, 2025. We focused on enabling instant booking at NYC bars and private dining rooms for birthday parties. We envisioned a mobile-friendly search engine for finding NYC bars and restaurants. One that displays private event spaces, packages, real-time pricing, and availability, all with instant booking.

We quickly validated the idea with friends who have also gone through the existing process: Submitting a form, receiving availability and a PDF with details over email, selecting the packages, and finally e-signing the contract. It seemed a no-brainer: consumers preferred the seamless and transparent approach. So, we scraped venues that were in trusted lists like Infatuation and Timeout, and what we knew was trendy. We created a landing page getthirdspaces.com, and an Instagram account. We were ready to go.

We based the initial market size on a fee-based pricing model for confirmed bookings. This was a fairly large opportunity in NYC alone. We realized our model depended on venues using the platform… without them, we couldn’t take a fee. So, we pivoted the initial search engine plan to instead focus on a tool that helps bars/restaurants manage their private events inquiries. Thus, they maintain the pricing and availability data we need for the consumer experience, and we can charge for confirmed bookings.

We started a spreadsheet to track outreach and began cold-calling bars near us. We also visited bars and became friends with a bartender who was kind enough to derail our research by giving us free drinks. We quickly learned our main competition was inaction, such as using email (maybe a form), phone, or walk-ins. Second, was Tripleseat, a heavy-weight event CRM platform. We also saw some newer companies like PerfectVenue and WaltzIn used.

From these conversations, we learned:

• No clear dissatisfaction among bars and restaurants: Tripleseat users seem satisfied, and those without software don’t feel underserved.
• Decision-makers are harder to reach than event managers: The buyer is usually the owner or management company.
• Lack of urgency for transparency: Bars and restaurants don’t publish pricing because they don’t see value in making it public.

We also spoke with a few folks who are working on similar event booking companies and learned:

• Corporate event demand remains soft: Revenue growth is slower than expected, largely because corporate events haven’t returned to pre-COVID levels. We reasoned this is the main revenue driver for these startups.
• High CAC from labor-intensive sales: Often due to the hands-on effort needed for sales and onboarding.
• Instant booking features aren’t used: Event coordinators avoid them because keeping pricing, packages, and availability updated takes constant effort. They also prefer to price each event manually to maximize revenue. For example, raising pricing when they spot a corporate email. Instant booking doesn’t support that kind of flexibility yet.
• Percentage-based pricing models don’t stick: Venues will bypass the platform to avoid booking fees, pushing these booking companies towards a monthly subscription model instead. Also, restaurants typically prefer their own payment systems, likely due to lower credit card processing fees, further reducing the opportunity to capture booking fees.

Despite the headwinds, we built and tested three prototype flows to demo to bars and restaurants. Through three iterations, we quickly learned that event booking is a messy inconsistent process. This led to two major decisions:

1. Solution approach: Simplify packaging + processes and re-educating owners, or supporting their existing, often inconsistent, processes.
2. Target customer: Start with either bars or resteraunts, since each has a different booking flow and sales cycle.

In the end, we made a third and final decision: to shut down ThirdSpaces.

Why we’re stopping

We found a problem. But:

• Not a big enough one. Coordinating events is neither frequent nor painful enough for venues to adopt dedicated software.
• Not a workable solution. Consumer experience depends on up-to-date venue data, but most venues aren’t incentivized to maintain the data. Booking flows also vary widely, which adds complexity to the product.
• Not a sustainable business. Restaurants and bars run on thin margins and often change ownership, making recurring revenue difficult to maintain. Meanwhile, high CAC from complex sales and onboarding further limits profitability.
• No clear product differentiation. ThirdSpaces doesn’t stand out. Competitors like TripleseatDirect and PerfectVenue Instantbook are ahead, and even they seem to struggle to gain traction.
• No clear early adopters. Most venues avoid early-stage tools. We were not able to find a venue to co-develop the product with from day one.

So, what did we learn?

First, marketplaces only work when both sides care. Consumers want speed in booking, simplicity in picking packages, and fairness in pricing. They’d happily use ThirdSpaces. But, venues aren’t incentivized to change and digitize, thus we don’t have the data for the consumers. Jumpstarting this marketplace is difficult. We respect how PerfectVenue is directly taking on Tripleseat’s CRM business to get the restaurants on their platform before they, one day, make their shift to instant booking.

Second, move fast to learn fast. Want to learn about a competitor? Request a demo and message them on LinkedIn. Want to talk to users? Reach out directly, show up where they are, or ask someone for an intro. Need a landing page? Pick a domain and launch it. Keep making forward progress. You’ll learn more and waste less time thinking.

Third and last, use AI sparingly for anything other than web searches and proofreading. Real learnings come from talking to people and connecting your own dots. AI won’t give you real user feedback. And generating endless prototypes is like pulling a slot machine. It’s addictive, but rarely useful. Be deliberate. Think clearly. That’s what actually drives progress.

What we’d do differently next time

• Start with what we know. If we don’t, find someone with firsthand knowledge who’s willing to work with us from day one.
• Find a way to secure a pilot customer before writing code. If you must, only write code that will enable you to get a pilot customer or user. It’s OK to be the user when you’re building for yourself.
• Keep checking whether you’re building a sustainable business. Some problems are real, but live in markets too small to support a business.

Final thoughts

• We experienced the consumer-side of this problem firsthand, and still believe there’s an opportunity for the right founder to solve for it.
• We’re grateful to everyone who shared time and insights with us.
• Working on restaurant tech, far from enterprise B2B, exposed us to a whole new set of challenges. That novelty was energizing. Solving real problems in new terrain reignited our excitement for building.

We’re walking away with lessons we couldn’t have gained any other way.

General Travel Tips

• Currency and Payment:
  Use a visa credit card without foreign transaction fees. The rate will match the unofficial exchange rate (see Blue Dollar), but cash often gets you up to 20% discounts. For safe cash exchanges in Buenos Aires, visit FAST CAMBIO - Currency Exchange in Recoleta.

• Reservations Are Key:
  For popular restaurants, tours, and activities, book ahead. High-demand spots (e.g., Michelin-star restaurants) get snapped up quickly—likely by other Americans doing the same thing.

• Airport Lounges:
  If using Priority Pass, pre-book lounges during layovers in Buenos Aires. They fill up fast, and otherwise, you’ll wait to get in.

• Learn Basic Spanish:
  While Buenos Aires is more English-friendly, much of the rest of the country is not. Knowing basic phrases makes the trip more enjoyable and helps you connect with locals.

Places to Visit

1. El Chaltén: A stunning mountain town in Patagonia and the starting point for incredible hikes, including to the iconic Mount Fitz Roy (the inspiration for the Patagonia logo).

2. El Calafate: A hub for visiting the Perito Moreno Glacier or driving to El Chaltén.

3. Mendoza: Argentina’s wine capital. Think of it as Sonoma, but with more steak and Malbec.

4. Buenos Aires: A beautiful, old-world city with European-inspired architecture. It’s worth visiting but keep your stay short (3 days max).

El Calafate

Overview

El Calafate itself is a small, quiet town, but it’s the gateway to the glacier and Patagonia adventures. Rent a car to make the most of your visit.

Where to Stay

• $$$ Posada Los Alamos: Cozy, lodge-like hotel with spacious rooms, included breakfast, a pool, sauna, and even a short golf course.
• $$$$ Explora en Torres del Paine: Luxurious and exclusive, though it books up fast. Seems nice if you can afford it, but I’d reccomend splurging in El Chaltén instead.

Restaurants

• Buenos Cruces Pasta Bar: Friendly staff, great pasta, small-town vibes. They learn your name when you enter. Book ahead for weekends or wait and eat late.
• La Cantina Piadineria & Beer House: Casual lunch spot with beer and empanadas. Great views of the water and mountains.
• Parrilla y Tenedor Libre Estilo Campo: Buffet-style restaurant serving authentic cordero al palo (Patagonian lamb). A local experience - no English spoken here so be prepared to order in Spanish!
• El Gaita Pizza Bar: Fine pizza but unexpectedly pricey.

Activities

1. Perito Moreno Glacier Hike: BOOK AHEAD with Hielo & Aventura (the only direct provider - others are just resellers). The Mini Trekking 2 tour is enough to enjoy this breathtaking glacier hike. They offer a longer tour called the Big Ice that we didn’t do. Afterward, visit the glacier’s viewing platforms and snack at the café. Drive early to enjoy scenic stops along the way.

2. Reserva Laguna Nimez: A peaceful bird sanctuary near town. Walk the guided paths, bring binoculars (or rent them there), and enjoy the natural scenery. Entry is ~$10.

3. Explore Lago Argentino: Drive north of town, stopping for photos, seeing the city sign, and to explore dirt roads along the water.

4. Glaciarium Patagonian Ice Museum: Surprisingly well-done museum on glaciers. A good pre-glaciervisit activity to get you excited for what you will see.

Drive from El Calafate to El Chaltén

• The drive takes way less time than Google Maps estimates because speed limits are flexible (locals often go 100+ km/h). Be cautious of guanacos (like you would for deer in the Northeast).
• Refuel before leaving El Calafate. El Chaltén has only one gas station, which gets crowded in mornings.

El Chaltén

Overview

A small, laid-back town surrounded by Patagonia’s majestic peaks. Many trailheads start here, so staying in town is ideal if you don’t have a car. Weather is unpredictable - layer up and be prepared for wind, rain, and sun all in one day.

Where to Stay

• \(\) Chaltén Camp: Glamping resort outside town nestled in a valley with beautiful views, incredible meals, and kind staff. The dirt-road drive to get there is picturesque and an adventure in itself.
• \(\)$ Explora en el Chaltén: A luxury lodge nestled in the national park. Expensive but reportedly worth it if you can afford it. They have some of the best trained guides in Argentina.

Restaurants

• La Cervecería Chaltén: Cozy spot for soup (Le Locro is amazing). We went back twice!
• The Asadores: Recommended by our Ice Trekking guide for traditional grilled meat.

Activities

1. Mt. Fitz Roy Through Hike:
   • Park at Sendero Laguna De Los Tres trailhead in town, then take a cab (coordinated by your hotel) into the park to hike back through the trail to your car. This hike is breathtaking and worth the effort. The through hike is also less crowded than the out-and-back so you can enjoy the isolation and quiet of the mountains.AllTrails Link.
2. Cerro Torre Viewpoint:
   • A lighter hike with gorgeous views. AllTrails Link.
3. Chorrillo del Salto:
   • A small waterfall a small way into the park. Skip the hike there and drive to save your legs for bigger trails.

Mendoza

Where to Stay

• In Town: Park Hyatt Mendoza—a well-priced, upscale option.
• South of Town (Luján de Cuyo): Entre Cielos Luxury Wine Hotel & Spa. We stayed here and loved the vineyard views and spa, but staff coordination wasn’t great. Plan activities yourself.
• Way South (Valle de Uco): Didn’t visit, but heard this area is stunning for luxury stays and wineries.

Restaurants

1. Casa El Enemigo Vigil:
   • Book ahead. Michelin-starred restaurant with incredible food and endless wine pairings. A must-visit!
2. Clos de Chacras:
   • Perfect for lunch during a wine tour. Outdoor seating near a koi pond and vineyard. Food was great; wine was average.
3. Entre Cielos Restaurant:
   • Easy if you’re staying at the hotel with surprisingly good food. Get the steak.

Activities

1. Horseback Riding in the Andes:
   • We visited Quebrada del Cóndor. Gorgeous wildflowers, streams, and pristine mountain views make it unforgettable.
2. Catena Zapata Wine Tour:
   • Argentina’s most famous winery. Book far in advance. Their “Blending Games” tour lets you create your own Malbec blend. Stunning setting with the Andes in the background. More Info.
3. Entre Cielos Hammam and Spa:
   • Pamper yourself with a luxurious massage and hammam experience. Details Here.

Buenos Aires

Where to Stay

• Recoleta: Stayed at Hotel Club Francés. Quiet, elegant neighborhood with an Upper East Side Manhattan vibe. Great for exploring nearby attractions.
• Palermo: Trendy area with Soho Manhattan vibes. Best for restaurants, nightlife, and energy. We’d stay here next time!
• Puerto Madero: Modern and upscale but isolated from the rest of the city.
• Montserrat: Grungier and more eclectic, but not ideal for accommodations.

Restaurants

1. Aramburu: Book ahead. A 2-Michelin-star, 18-course tasting menu. Expensive for Argentina but cheap compared to the US. It’s a once-in-a-lifetime experience. Just book it!

2. Florería Atlántico: Rated on the top 100 bars in the world.A speakeasy under a flower shop. Excellent cocktails, world-class vibe. The drinks are so cheap that you are certified to leave buzzed.

3. Chori: Casual spot for delicious chorizo sandwiches in Palermo. Might be a little too toursity since it’s write up in the New York Times.

4. Rapanui: Buy fancy chocolates for gifts or for yourself!

Activities

• Tango Show at El Querandí: Dinner and tango performance. Great tango, questionable dinner, lots of wine. Read a review that said the tango brought them to tears!
• Recoleta Cemetery: Wander the elaborate tombs of the wealthy elite. Eva Perón is buried here. This is probably a place where a tour guide would make it more interesting. We tried to use ChatGPT once there, but it didn’t know a lot about the history.
• Cooking Class (Empanadas): Fun group activity to learn local cuisine and meet fellow travelers.
• Ecoparque Buenos Aires: Explore the abandoned zoo, now overrun by nature and turned into a public park. See some animals roam freely and interact. Other larger animals are still in their cages because they’re too big or old to move. It’s a nice way to spend a few hours.

The first thought you might have when you decide to make a blog is “wow I have a big ego”. Your second thought might be “what do I have to write about”.

Well, if your answer is nothing then you can write about how you made a cheap blog. That is what I’ve done. This blog post is about hosting your own cheap blog.

🎯 Prerequisites

1. Github and AWS account
2. Some familiarity with coding and your terminal
3. Homebrew Package Manager

💸 Cost of blogging by platform

At the time of writing (Jan 2021) here are the prices I found online to use your own custom domain (i.e. blancotech.com). Ignoring the price to maintain the domain (I use go daddy and it’s around $20/year)

To be honest, I did this price analysis after and didn’t know Medium was free. Also, I didn’t think to use GitHub. So, do what you like 🤷‍♂️.

⚙️ Setting up Jekyll

Jekyll is a static website and blog framework. It’s easy to get up and running. You can choose custom templates for your blog and easily create new posts using markdown syntax.

I’m using a MacBook on Mac OS Big Sur v11.1. I first tried to follow the instructions on Jekyll’s site https://jekyllrb.com and input them into my terminal

$ gem install bundler jekyll

but this immediately errored

$ gem install bundler jekyll
Fetching bundler-2.2.4.gem
ERROR:  While executing gem ... (Gem::FilePermissionError)
You don't have write permissions for the /Library/Ruby/Gems/2.6.0 directory.

I determined it was because I was using the ruby that is native on my Mac OS. To confirm this I checked what my system was using

$ which ruby
/usr/bin/ruby
$ which gem
/usr/bin/gem

To install the latest version of Ruby I used Homebrew

$ brew install ruby

Next, I told my shell environment to use the ruby installed by homebrew. By adding the path to my shell file ~/.zshrc.

$ echo 'export PATH="/usr/local/opt/ruby/bin:/usr/local/lib/ruby/gems/2.7.0/bin:$PATH"' >> ~/.zshrc

Then I refreshed my terminal with the file using $ source ~/.zshrc. Alternatively, you can open a new terminal window and this will do the same thing.

⚠️ If you’re not using ohMyZsh then replace the ~/.zshrc in the above command with ~/.bash_profile.

To confirm that you’re using the brew installed ruby run

$ which ruby
/usr/local/opt/ruby/bin/ruby
$ which gem
/usr/local/opt/ruby/bin/gem

Then running $ gem install bundler jekyll worked 🎉!

🛠 Create your project

To create your project go to the directory you want to create it in. Next, run jekyll new myblog. You can replace myblog with whatever name you’d like to give it.

Run jekyll serve to launch your site and view it in your browser at the URL localhost:4000.

🌐 How to get your code into a repository

In order to get your site online, you’ll need to put your code into a repository. For this, I used GitHub.

1. Go to your GitHub and create a new repository with nothing added to it (no git ignore, etc).
2. Setup the repository by running git init in your project’s directory in your terminal
3. Add the files you will commit by running git add . in your terminal
4. Commit the changed files by runninggit commit -m "My first commit" in your terminal
5. Tell your repository where to push the code by copying the repositories link from GitHub then running $ git remote add origin <your_link> in your terminal. Your link will look something like git@github.com:YourUserName/yourlink.git
6. Finally, push the code by running git push -u origin master

🌐 How to get your site online

So the blog was created! To host it online for the world to access it use AWS Amplify and follow the tutorial here: Deploy your Jekyll Site using AWS Amplify — with only a few clicks.

You now have a blog using Jekyll hosted on AWS Amplify 🍻.

Travel Up from New York City

• Lunch in Portland, Maine: Stop at Duckfat for incredible poutine.
• Scenic Route: After Portland, skip I-95 all the way. Instead, take I-295, then switch to US-1 near Brunswick for a picturesque coastal drive. Adds about an hour but worth it.

Where to Stay

• Short Trips: Stay near Bar Harbor. It’s the main hub with restaurants and nightlife, and you can walk back after drinks.
• Longer Trips: Anywhere on the island works.
  • We stayed near Southwest Harbor, which felt more authentically “Maine” with fewer crowds.
  • Perks: Quiet area, fire pit, tasty lobster shack nearby, and close to Seawall Campground for scenic dinners.

Town Vibes

• Bar Harbor: Main tourist hub with lots of restaurants, nightlife, and close to major hikes (10-20 min drive).
• Northeast Harbor: Upscale vibe with fancy homes; feels like the Hamptons.
• Southwest Harbor: Quaint and quiet. Great if you want peace, but it’s 30-40 minutes from the main hikes.

Top Hikes (Ranked)

1. Beehive Trail
   • Iconic hike with metal rungs for climbing—great adrenaline rush! Includes a beautiful pond-side trail.
2. Sand Beach
   • Near Beehive, one of Maine’s rare sand beaches. Perfect for a post-hike stroll.
3. Beech Cliff Trail (via Beech Cliff Loop)
   • Quick 1.5-mile hike with steep ladders and stunning views at the top. Great for a lunch break.
4. Gorham Mountain
   • Easier than Beehive but just as fun. Offers ocean views and optional adventurous cliff-side detours.
5. Thunder Hole
   • Dramatic rock formation with splashy waves. Timing is key; it wasn’t very loud when we visited.
6. Bar Island Walk
   • During low tide, walk across the sandbar to Bar Island. The island’s hike is average, but the sandbar experience is unique.
7. Jordan Pond + Bubbles
   • Popular but crowded. Scenic views and a chance to try popovers at Jordan Pond House (if it’s not too busy).

Restaurants

• Charlotte’s Lobster Pound
  • Delicious lobster rolls with a bonus: goats in the back! Pick up food and eat at Seawall picnic benches for water views.
• Eat-a-Pita
  • Great breakfast spot near Southwest Harbor with surprisingly top-notch service.
• Café This Way
  • Located in Bar Harbor. Don’t miss their breakfast if you’re hungry.
• The Reading Room
  • A “fancier” spot in Bar Harbor. Decent food, NYC-level prices, and a romantic outdoor space overlooking the harbor.
• Mt. Desert Island Ice Cream
  • Thick, dense ice cream that absolutely hits the spot.

Other Activities

• Sunrise at Cadillac Mountain
  • Classic Acadia experience. Drive to the peak for stunning views. Note: Early wake-up can wreck your day from exhaustion!
• Bioluminescent Kayaking
  • Magical and worth the 1.2-hour drive. Highly recommend this experience!
  • Castine Kayak Bioluminescence Trips
• Climbing Adventure
  • Challenging but exhilarating. Perfect for thrill-seekers!
  • Acadia Mountain Guides
• Sunset Sail
  • Cheesy but worth it—the best sunset I’ve ever seen. Drinks on board make it even better.
  • Downeast Windjammer Sunset Sail