Matt Gaiser

I Shipped a Merge Request to a Repo I Had Never Seen, in a Language I Have Never Used

Matthew Gaiser — Sun, 20 Sep 2026 12:00:00 GMT

GitLab ran a hackathon in June around Orbit, their new knowledge graph for code. The rule for the Contribute track was pretty simple: the first forty merge requests merged into the Orbit codebase with the right label before the deadline would win a prize.

I do not write Rust, which is what almost the entire codebase is written in. I had never opened it, did not have a Rust toolchain installed on my laptop, and had never run their build system or read their contribution guide.

I got a merge request merged anyway, and it took an afternoon. The results came out in September, and I am on the Contribute Track winners list. Twenty six people merged sixty one changes into Orbit over the two weeks. Nineteen of us got cash prizes, as much for getting there early as for doing good work.

I did not write the code. I am not sure I could have. I pointed Claude Code at the project and told it to find me a contribution and ship it.

The code was the easy part, and it was not the point

The change was pretty small, a command called orbit list that shows which repositories you have indexed locally. About two hundred lines including tests. A competent Rust developer would not have found the ticket particularly impressive.

Writing that code was maybe a tenth of the work. The rest was the boring stuff that usually stops me from contributing to a project like this.

You have to read through the issue tracker and find something you can finish that someone actually wants. Then there is the toolchain to install, the build commands to learn, and the house style for comments and commit messages to figure out. Even the tests have to be written the way this project writes tests. All the linters have to pass before a human will look at it, and after that you still have to get through review without annoying the maintainers.

Getting through all that is the expensive part of contributing to a large project you do not already work on. I think it is why most people never send the patch. The agent handled it all.

The interesting failure was a wall, not a bug

My first attempt failed for a reason I would not have figured out on my own.

I pushed the branch from a personal fork and opened the merge request, and the pipeline went red. The tests all passed, so there was nothing wrong with the code. It turned out that personal forks on GitLab do not get the runners and the dependency proxy that the full test suite needs. I had working code and no way to get it through CI there.

The agent read the failed pipeline logs and worked out that the problem was with the setup. It also found that GitLab runs a community fork program for exactly this problem. Getting in required an access request and approval from a human, which took a few hours. Once that came through, the agent pushed to the community fork, opened a fresh merge request, and closed the old one with a note explaining why. The full suite passed there.

There was no code to fix here. The useful part was being able to read the error log and find out that a solution already existed. A lot of contributors hit the same red pipeline and got stuck without ever figuring out why.

The robots reviewed the robot

GitLab’s own AI reviewer, Duo, read the merge request and left two recommendations. A second review bot approved it with a couple of small notes. The agent fixed everything the two of them raised and replied to each thread, all within the hour. A human maintainer then approved it and merged it the same day.

So I had a bot write the change, two other bots review it, and the first bot fix what they flagged. After all that, a human decided it was worth merging. That was it. It bounced from bot to bot and only at the end did a human take a look.

What this changes

It is easy to forget how much would have been in the way a few years ago. I would have spent a weekend getting Rust installed and the project to build, given up somewhere around the third failing linter, and never sent anything. Between the unfamiliar language, the build system, the conventions, the CI quirks, the fork logistics, and the review etiquette, there was plenty to get stuck on.

I had no Rust, no toolchain, and no idea what this codebase even was. A few hours later, there was a change merged into GitLab with my name on it.

On Spotify: My Interview with The Solver's Edge

Matthew Gaiser — Thu, 22 Sep 2022 17:08:38 GMT

My interview on The Solver's Edge podcast, available on Spotify.

On Dev.to: Free Code Coverage for Pull Requests on GitHub

Matthew Gaiser — Sat, 28 Aug 2021 20:34:03 GMT

How to add free code coverage reports to your GitHub pull requests. Published on DEV.

Who will clean the toilets?

Matthew Gaiser — Sat, 21 Aug 2021 18:25:30 GMT

Discussions about basic income and sometimes immigration devolve into a discussion about how the unpleasant work of society will be done if coercion of some form is not required for people to do work. The phrase I have most frequently seen used is “who will clean the toilets?”, which I have mostly encountered on Hacker News and Reddit.

The assumption in this statement is that toilet cleaning is a required activity and we will need to be spending as much time cleaning toilets in the future as we currently do. While I admit that a certain amount of toilet cleaning will likely always be required, toilets, as they are designed, are very hostile to sanitation.

Consider this public washroom in Calgary that inspired this blog post. The toilet has a lot of surfaces, hard edges, and places that urine and splashing poop flakes could hide.

Or consider the hinges.

Under the toilet was difficult to even get a photograph of. I was kneeling to just get the bowl. Luckily the floor was clean in that case.

Or consider the shelf in the washroom. While an excellent idea to give people a place to put a coat, phone, or purse, the shelf design of a glass plate on top of the base shelf leads to a place where garbage, dust, and gunk can accumulate.

I struggle to believe that if the designers of that washroom and its components were tasked with cleaning it, they would have designed any element of the washroom the way that they did. Nothing about the design accommodated easy cleaning.

And this was a mall bathroom, where most patrons are clean, sober, and not really intent on mischief. A bathroom in a stadium, a university, or a school would require even more cleaning given who the washrooms serve.

Now, many will argue that toilet cleaning is not meant to be taken as a literal task, but rather as a representation of the many unpleasant tasks that exist, but again, how many of those tasks are only done that way because human labour is so cheap that nobody needs to consider ways to conserve it?

So, who will clean the toilets? Let’s first see if toilets really need so many people spending so much time cleaning them.

On Dev.To: The Chrome Extension Deployment Checklist

Matthew Gaiser — Sat, 01 May 2021 08:51:21 GMT

Everything to sort out before you publish a Chrome extension, as a checklist. Published on DEV.

On Dev.to: How to debug an Office.js Manifest File

Matthew Gaiser — Tue, 22 Dec 2020 08:56:12 GMT

How to track down errors in an Office.js add-in manifest file. Published on DEV.

On Stack Overflow: How to communicate more deliberately when working remotely

Matthew Gaiser — Thu, 22 Oct 2020 11:21:56 GMT

How to communicate more deliberately and efficiently when your team works remotely. Published on the Stack Overflow Blog.

On Freedom With Writing: The Anatomy of an Instructional Pitch.

Matthew Gaiser — Fri, 21 Aug 2020 17:49:34 GMT

How I pitched an instructional article and got paid $500 for it, broken down step by step. Published on Freedom With Writing.

There Must be Useful Processes For Resolving Real World Edge Cases

Matthew Gaiser — Thu, 06 Aug 2020 05:39:32 GMT

One of the tragic facts of the Beirut Explosion is that not only was it preventable, but that the problem was reasonably well known. Many people knew that there was a bomb in the harbour and many tried to warn people about it, but nobody was sure who was responsible nor who should be demanded to take action.

The world is full of cases like this.

Most of us have worked in a company doing something wasteful with no clear path to resolve it. Entire departments can be generally aware of the waste, yet have no idea what to do about it.

On large government projects, plenty of people know when a project is going off the rails. But what are they supposed to do? It is not clear.

In most organizations, the answer would be to “tell your manager.” But generally they do not know so they pass it up to the next manager and then on to the next manager and this game of telephone continues until some manager forgets about it or does not know who to pass it on to.

Some organizations attempt to use hotlines, but these are not the answer, or at least not the entire answer. Wells Fargo had hotlines for reporting fraud, but whoever was operating those hotlines either had no power to change anything or was actively participating in the fraud (or was at least managed by someone who was participating). Employees cannot have meaningful confidence in that kind of system.

Even if the hotlines are not honey pots to catch disgruntled employees, it is equally possible for them to exist merely as morale boosters, with no avenue for actually resolving problems.

Maybe companies and governments just need a Department of Edge Cases. A group that has the highest level of authority to handle all of those issues which don’t clearly fall under anyone’s mandate. It is not as though large organizations face a shortage of problems internally for this group to tackle and they could very cheaply prevent some very serious issues.

There Should Be Social Engineering Bug Bounties

Matthew Gaiser — Sat, 01 Aug 2020 04:11:48 GMT

Millions of dollars are available if you find technical flaws in major products that allow you to breach their systems.

How much can you get if you can use a social engineering hack to breach a system?

Generally nothing.

A Google search for “social engineering bug bounty” returns a list of programs that explicitly exclude any social engineering breach as being eligible for a reward.

At first glance this makes sense. Figuring out how to trick a human is hardly a consistent and repeatable process, so fixing it is not an easy or permanent fix. Social engineers could easily exploit the same bug over and over again.

However, whether it be most ransomware attacks or the recent Twitter hack, the vulnerable point was not in the technology, but in the human operating it. The hack was in convincing a user authorized to take administrative actions to unknowingly allow the hacker into the system. The hacks into the 2016 Democratic Presidential Campaign were also social engineering hacks.

Social engineering is not a minor attack type, but rather the mechanism behind some of the most well-known hacks in recent history.

The main argument against this is that it would rapidly get expensive because so many people would give away their password, leak important information, or click a suspicious link. I could easily see organizations losing tens of thousands of dollars in the first few months of this program.

But if a social engineering bug bounty would be extremely expensive to a company, they have a serious problem.

If you have a server on the internet, you know that it is under constant attack from automated bots pinging ports and checking to see if you left your WordPress credentials set to their defaults.

The people in organizations are equally under attack. Indeed, they are some of the most vulnerable parts of the system because nobody is routinely updating the people on new attack strategies. How many of them know that they could be targeted? How many of them could be fooled by a carefully crafted phishing email? People who can be fooled by phishing emails are essentially default admin passwords.

A technical bug that allowed an outside person to gain access to an admin account would be worth thousands, if not tens of thousands of dollars.

Gaining that same access through social engineering and reporting it to the company would get you a visit from the local or perhaps even federal police.

It shouldn’t be this way.

Companies should establish social engineering bug bounties. If you successfully obtain the Office 365 credentials of an employee, that should earn you perhaps $200. A Gitlab login or a more important login might be worth $900.

It need not be a free-for-all that leads to mass employee harassment. Social engineers could be required to register beforehand, be rate-limited in how many emails/calls they could make, and certain approaches (such as pretending to be a family member in distress) could be ineligible for rewards. They could be required to demonstrate that the credentials were obtained through some kind of social interaction rather than from some online data dump. Precise documentation of every source of phishing information could also be demanded.

Social engineering has proven its power as an attack method, especially when applied creatively. Bug bounty programs should reflect the potential impact of that.

You Should Spend 250 Hours Learning About Mortgage Interest

Matthew Gaiser — Fri, 03 Jul 2020 05:25:36 GMT

People have a very casual attitude to one of the most consequential financial decisions of their lives, their mortgage interest.

Let’s assume you are buying the median Canadian home for the median Canadian family income. Median house is $531,000. Median family income is $92,700.

Putting 20% down, you have a mortgage of $424,800.

Let’s consider two scenarios. A 1.99% interest rate (generally what you can get right now if you compare providers) and a 2.25% interest rate (what people on this subreddit seem to be getting). We will assume a 5 year term and a 25 year amortization.

At 1.99%, you will pay $38,787.28 in your first 5 year term.

At 2.25%, you will pay $43,944.17 in your first 5 year term.

The difference in interest paid is $5,156.89‬, just over the first term. The lifetime difference is $130,344.06 – $114,231.48 = 16,112.58‬.

Assuming that the $92,700 is a dual income family, let’s assume it is $75,000 after taxes from two people working a total of 70 hours a week. 70 * 52 = 3640. Thats $20.60 an hour after tax. That is what Mr. and Ms. Median value their time in the workforce.

5156.89/20.60 = 250 hours. 4 * 70 is 280, so it is just about equivalent to a month of work.

The difference between 1.99% and 2.25% is a month of salary for a median family.

This is not to say that you should buy exclusively based on the rate as things like portability and prepayment certainly matter, but it is worth spending a large chunk of time sitting down and figuring out mortgages inside and out.

From an effort perspective, you should theoretically be willing to spend the equivalent of 250 hours (a month of full time work) understanding your mortgage and getting a lower rate.

Would you make 10 phone calls to get $5000? Would you scour the internet for hours to get $5000? Would you aggressively negotiate to get $5000? Would you spend time reading a book for a $5000 bonus at work?

As that is what is at stake. 0.25% seems so tiny and harmless. It looks like a small number, but it isn’t.

On Software Engineering Stack Exchange: Does scrum ruin great engineers or are you doing it wrong?

Matthew Gaiser — Tue, 30 Jun 2020 07:22:25 GMT

Does scrum ruin great engineers, or are teams just doing it wrong? Published on the Stack Overflow Blog.

On Opportunity Desk: I was Named Opportunity Desk's June 2020 Young Person Of the Month

Matthew Gaiser — Tue, 02 Jun 2020 08:44:30 GMT

Opportunity Desk named me its Young Person of the Month for June 2020. Read the feature on their site.

COVID Has Shown that Elder Care Needs Disruption

Matthew Gaiser — Tue, 02 Jun 2020 08:19:01 GMT

The coronavirus pandemic has revealed several industries that are in need of substantial disruption.

Right at the top is the eldercare industry, specifically retirement homes. Whether it be in Spain, the United States, or Canada, many retirement homes have proven themselves to be unreliable guardians of the elderly. Thousands have died or been abandoned due to the incompetence and/or indifference of care home companies.

However, the problems with elder care facilities long preceded COVID. They are often not really accountable to anyone in practice. They are heavily dependent on poorly paid labour which frequently turns over. They do a poor job of solving the other challenges the elderly face, such as loneliness and elder abuse. In addition, elder care is expensive. Most retirement home facilities are difficult for many families to afford.

And this is all while the population of elderly people is relatively low and the available resources can be spent on fewer people. The number of elderly people is set to double over the next few decades in the United States and Canada, only raising the economic costs.

Specific Areas of Innovation

If you are interested in tackling this problem, here are some areas you might consider as starting points.

Accountability

Retirement homes have few accountability mechanisms. Government inspection departments are understaffed and often not empowered to proactively investigate. The elderly in the homes lack access to technology that would allow them to contact the appropriate investigators, even if they were comfortable using it. Many elderly are reluctant to burden their kids with issues so they would rather suffer than make them worry. Retirement home margins are often low, which encourages a nickel and dime mentality and relentless pressure to drive costs downward.

Given how coronavirus has exposed the poor operations and standards of these facilities, a way is needed to verify that retirement homes are providing a quality product and if they are not, rapidly expose that to the residents, their family members, or the government.

Socialization

Every nursing home has stories of residents who are only visited a couple of times a year. Depending on the study, somewhere between 18% and 40% of nursing home residents consider themselves often or always lonely. While technology has a poor track record of actually reducing loneliness, it may be better than nothing here.

Automation

This is the obvious one, but there could easily be some low hanging fruit. Those workers at the retirement homes are working at a rapid pace at something. Could that something be automated in any way?

Aging in Place

Considering how the elderly have been treated by retirement facilities, I can see far greater reluctance to use them and far greater interest in an alternative. Are there ways technology could help people stay in their homes?

Labour Multiplication

The industry is heavily dependent on minimum wage human labour and it can just barely find enough workers. It is chronically accused of understaffing while structuring shifts so that workers need to work at multiple care facilities to make a living (a possible contributor to the spread of COVID). There is no room to cut costs further under the existing model.

Even if large parts of the elder care process cannot be automated, perhaps technology could allow personal care workers to achieve more within their shift.

Amazon Should Not Be the Easiest Website to Search for Shoes

Matthew Gaiser — Fri, 22 May 2020 06:11:11 GMT

This pandemic has allowed me to spend days not wearing shoes and when now I put my current pair on, I now realize that they have worn out and are no longer comfortable. It is time to purchase a new pair.

I have some slightly special shoe needs. My feet are wide, so I need a shoe specifically designed with that in mind. I also like long laces because I stretch my feet in the shoes and if the laces are too short they frequently come undone.

Otherwise, I am not picky about what I purchase. The shoe can be a bland grey or a bright orange. I am not loyal to any particular brand. If I can get an excellent shoe, I am willing to pay more. All I need is a general walking/running shoe for day to day usage. But those two things are non-negotiable. Of all the people I know, I have the fewest requirements for shoes.

These are not rare things to find in shoes. Most shoes have sufficiently long laces. Go into any shoe store and you will find a few wide versions scattered around the shelves for most shoe models. Before the internet, I had no trouble walking into a store and finding a plausible option within a size of what my ideal size would be.

I usually would go straight to Amazon to purchase anything, but doubted that they could do shoes and apparel better than retailers who dedicated vast amounts of their efforts to selling shoes. So in addition to Amazon I decided to check out Sport Chek and Foot Locker, which are two large sellers of shoes in Canada.

My search did not start off well. Sport Chek’s website offers me no way to search for shoes based on the width and does not even show that information on the page about the shoe. Having repeatedly purchased shoes from Sport Chek, I know they are available but I cannot search for them or even identify which models do have wide sizes by reading through the shoe profiles. Even worse, after a few minutes my session timed out on the website and it sent me away from the product page to the home page.

The page on Sport Chek timing out. Clicking “continue” sends you to the home page.

Foot Locker lets you search by width, but just one width.

Footlocker does allow me to sort by width, but only select from a single width, the standard one. They show me 327 types of shoes with the option to search by width, but they do not seem to have loaded those wide options into the computer system.

Having spent a half hour hunting for information about shoe width at two of the largest shoe retailers in Canada, I could find no way to identify whether a shoe was available in a size that would fit my foot.

Going to Amazon.ca, once I entered the shoe category I was greeted with a “Men’s Shoe Width” selector just slightly down the page. It took less than 30 seconds for me to filter for wider shoes. Within 10 minutes I found one that met all my requirements and even allowed me to select from several different colours.

The Width Search on Amazon.ca

It is absurd that it is easier to buy shoes on Amazon than even at dedicated shoe retailers online, but that is how it is.

Amazon specializes in nothing but behaves more like a specialty retailer than the retailers which actually focus on selling these items.

Things You Can Do If You Do Not Have a Tech Internship This Summer

Matthew Gaiser — Mon, 04 May 2020 19:08:42 GMT

So you didn’t get an internship this summer or had yours rescinded. Fortunately, there are plenty of other options for productive things you can do which will increase your skills and boost your resume at the same time. Some will even let you earn some money directly. Others just boost your earnings down the road.

Hackathons

With developers trapped inside their homes, they have turned to organizing numerous online competitions for each other. That means there are more hackathons than I have ever seen active and open to developers all around the world.

They range from ones targeted at teenagers to $100,000 challenges meant for teams of professionals, so there is something for everyone at every skill level. Many also include workshops to help you learn new skills, often in areas you never before considered.

As a bonus, the prizes aren’t bad. You might be able to win a couple hundred dollars if you have a good project or perhaps a cool new drone.

The best sites to find hackathons include:

DevPost: https://devpost.com/hackathons
DevPost is a site for hosting hackathons and consequently, many are listed there. Probably the main website for hackathons.

HackerEarth: https://www.hackerearth.com/challenges/hackathon/
HackerEarth is a DevPost competitor. They do not have as many events and the user experience can be frustrating at times, but they sometimes have some very good events as they are more corporate-focused.

Major League Hacking: https://mlh.io/seasons/na-2020/events
MLH is a hackathon league which is organizing a swath of events for the summer. It is very beginner-friendly and targeted heavily at students. They also run many recruiting events throughout the year.

Innovation Challenges

An innovation challenge is essentially a competition to find the best ideas to solve a given problem.

Innovation challenges also usually pay winners prize money. It is not always a lot, but sustained efforts can add up quickly.

My favourite sites to find innovation challenges:

MindSumo: https://www.mindsumo.com/
MindSumo is my all time favourite for innovation contests and I have done hundreds of challenges with them. They usually have a new one every week and it could be anything from developing a new food product to designing new technology for outer space.

Another great thing about them for beginners is that prizes are awards for the top 50% of entries. 1 in every 2 people will win some prize money. Even if you don’t do so well the first few times, you will probably still earn a few dollars.

Eyeka: https://www.eyeka.com/
Eyeka contests are mostly geared towards people who can create stunning visual demonstrations of their ideas as every solution must be presented with a high-resolution photograph. If graphic design is a side hobby of yours (as it is for many frontend developers) this is a great site for you.

Even if you are not great with the visuals, I have won a few of these using little more than Microsoft Paint, so do not give up hope.

HeroX: https://herox.com/crowdsourcing-projects/enter
Most of these are not that accessible to the average student or if they are, they are mostly for large teams, but there are still some there are you can enter on your own with just a few hours of effort. It is even better if you are the kind of person who can put together a team, as then many of the challenges will be open to you.

Agorize: https://www.agorize.com/en
Agorize is geared towards competitive rounds of competitions. Instead of submitting an idea and waiting for the judges to issue a final recommendation, there are instead rounds of judging with ever-increasing submission requirements. You also get to work with mentors to refine your projects.

Whether you choose to use it depends on whether you prefer short projects or longer more enduring ones. You also need a friend to work with on these as the typical competition requires at least two people.

Interview Preparation

Technology is one of those fields where the skills tested in interviews are not skills you will pick up building projects or even working in anything but the largest technology companies. I have repeatedly had to demonstrate my ability to write some kind of sort. However, when is the last time you wrote a quicksort outside an algorithms class or an interview?

There is one classic book for this. Cracking the Coding Interview.

If you stay in software development, you will need concepts in it for every interview you have from now until our field pivots to something new (which is unpredictable in software development). This may be a good time to build those fundamental interview skills so that you can be ready to go when the world re-opens.

You never know when you will have the opportunity to interview for a great company and need your algorithms skills. I blew an interview for a FAANG simply because I didn’t keep my skills in this area fresh, had only 2 days to prepare, and skipped over studying the content actually in the interview. Don’t be like me.

Found a Startup

If you scrolled all the way to the bottom, then the many other links were not interesting or challenging enough for you. So why not take on the ultimate challenge of finding your own challenge and solving it?

As cliche as this answer is, there is a certain merit to it. Numerous billion-dollar companies were founded in the depths of the last recession. Coronavirus could very well reshape society, which will clearly provide certain opportunities. Governments are also likely to invest heavily in innovation to get the economy going again.

This is the one thing on this list which I have not done extensively myself, so I will just point you to some Google Cloud and Azure credits and let you take things from here. Paul Graham’s essay on startups is also a popular starting point.

If you need an idea for a problem to solve, check out the official Y Combinator list of desired startups. You can also comment below on your areas of interest and I can help you find an idea. If there is some resource you need, I am also a very good Google sleuth.

How to Win at MindSumo from the #1 All-Time Solver

Matthew Gaiser — Sun, 03 May 2020 20:55:02 GMT

The QuickStart Guide to Winning at MindSumo

You are probably here because, like many people, you have asked me for advice on how to be better at winning MindSumo challenges. I have finally gotten around to putting together a guide for doing this.

For anyone who may have somehow stumbled across this accidentally, MindSumo is a crowdsourcing innovation website where companies post challenges and people can submit their solutions. The companies rate the solutions and depending on how you score and how many people enter, you can be awarded prize money.

I am the #1 ranked solver on the website, in challenges won, prize money won, and challenges entered.

Here is how I do it.

**Think Beyond the Written Question(s) and Challenge Assumptions
**
Most MindSumo solvers are students and they are translating that experience with questions over to MindSumo. With schoolwork, particularly on tests, questions are comprehensive and written by someone who knows the correct answer. There is no expectation that your answer to address information not requested in the question and there is little room to challenge the question as an incorrect framing of the problem or just outright flawed in other ways.

On MindSumo, the people asking the question do not know the correct answer. That is why they are asking you and putting up a $1600 prize pool for solutions. Before they read the entries, they do not know what the best answer will be. There is no answer key as MindSumo is not an exam but instead an exploration., one where you are free to break free of the rules you might use for writing a test.

Some of my best entries have started out with something like “The questions asked do not address the real challenge”. If some question asked in the challenge does not make sense, feel free to state that in your solution as long as you include an explanation as to why. If you can successfully argue that one of the questions is flawed, you immediately put yourself above the other solutions based on that flawed information.

In one challenge, they asked for information on how to target a certain market. I argued that the target market was completely wrong and argued for a different market. In another challenge, I argued that the business model of the entire company is wrong and that they are asking the wrong question while spending most of the text suggesting a better business model.

I won both

Determine Their Motivation for Posting the Challenge

MindSumo challenges are not free, so the company decided that it was worth spending several thousand dollars asking the world that question. Why? Being able to answer that question goes a long way toward producing a winning entry.

For some companies, it is about a lack of domain expertise. They actually do not know a lot about the area they want solutions for. For others, it is just an alternative form of market research. They want to see how customers think. Most seem to want ideas for new products and services that they can offer customers.

You can identify their motivations by taking a look at the company as well as the questions that they are asking in the challenge itself. For example, this challenge is clearly partially market research because they are asking every participant to rank various preferences of theirs as part of a survey.

What solvers should do for this challenge is not only identify an ingredient, but also provide support for why that ingredient could be popular or why it is already popular.

Considering challenger motivations in that way lets you show why your answer is relevant to their needs and that helps you to earn a higher score.

Take Innovations from Elsewhere and Apply Them to the Challenge

If you already have an original idea then you can probably skip over this section, but many people get stuck on thinking of an idea. When I cannot think of a good idea, I take my inspiration from how a similar problem has been solved in another industry.

One of my solutions came when I was bored in class and decided to unscrew a ball-point pen. The challenge was to find a new solution for lubricating things and I used the ball-point concept as the core of my solution. I am sure everyone who entered the challenge and everyone who rated the challenge knew what a ball-point pen was and how it worked.

Your solutions do not need to be truly unique. They can just be well-known ideas applied in a unique way. That ball-point pen concept won 1st place in that challenge simply because it was such an easy solution that nobody had thought of using in that way.

Feel Free to Make a Wild Proposal. The Company Probably Knows the Obvious

Remember, the company paid MindSumo (and hopefully will be paying you) for this solution. That means they will have at least done a Google search to determine if their question can be answered that way. Anything obvious and hyper-realistic would have already been covered by the company internally.

You must provide something they have not seen to win

Write Clearly and Concisely and Use Illustrations

The average MindSumo challenge receives some 100-300 solutions. Each solution is roughly 300 words, which makes the total reading for each rater equivalent to a small or medium-sized book. I did rating when I worked at ATB Financial and it took days as we received over 400 entries.

With so many to read, entries that do not rapidly make clear what the idea is might get skimmed and given an average score. Entries that are hard to understand may be given an average score. The best entries can be basically understood within 15 seconds of skimming.

Keep Solving and Keep Trying

I entered my first two challenges about 7 months before I started doing MindSumo regularly and I lost both. I only ended up returning because I had a month of free time that I needed to use. Without that free time block, someone else would have been the #1 ranked MindSumo solver.

You probably will not win your first few challenges. That is ok as it takes time to really understand the platform.

Just keep on trying. You cannot win them all.

On The Overflow: How to Have Better Meetings

Matthew Gaiser — Thu, 30 Apr 2020 23:59:22 GMT

Practical ways to run better meetings, in person or remote. Published on the Stack Overflow Blog.

On The Overflow: How to Onboard Yourself to Your New Job.

Matthew Gaiser — Mon, 24 Feb 2020 00:11:13 GMT

How to onboard yourself to a new job when your employer doesn't. Published on the Stack Overflow Blog.

I Was Named One of Major League Hacking's Top Hackers of 2020

Matthew Gaiser — Thu, 20 Feb 2020 02:56:43 GMT

Major League Hacking named me one of its Top 50 Hackers of 2020. See my profile on the MLH site.