Shahriar Haque

O you, living at the mercy of desire

Under the weight of expectations, why do you tire

Close by, the wind that sways and stirs me

It knows all, it’s there to urge me

Don’t suffocate the depth within you

Nurture your soul, don’t be afraid to grant yourself what’s due

I’m assailable, but I’m in my mind

I don’t need you telling me that I’m starry-eyed

I don’t wanna hold up if I can’t break

I don’t wanna shut down just to feel safe

I’m pathetic, so frenetically romanticizing

But there’s just something so poetic in the clutter, you can’t wear me down

Oh, you can’t wear me down

– Lyrics by Abdullah Siddiqui, Maanu & Xulfi

A few days ago, I shared a post on reddit about AI art depicting women in rural Bangladesh. Today I want to share my own experience of living in rural Australia for over 4 years.

Firstly, a note about demographics. In Sydney, depending on the neighborhood, you can walk around all day and never meet a white person. However, rural Australia is predominantly white.

The second thing I noticed since day one - women work literally everywhere, not just in specific industries. I worked at a petrol station for a few years and we have always had women drive huge fuel-delivery trucks and single-handedly manage the plumbing to load fuel into our tanks. It’s not uncommon to see women work in road construction too. I used to see them operate drill machines and drive forklifts to clear up rubbish. If you ever visit a hardware store here, you’ll find women showing customers how to operate chainsaws, lawnmowers etc.

Thirdly, even in family-run businesses like farms and workshops, its the women who manage all the financial matters. Men typically handle the grunt work while the wife or daughter handles everything from customer service, cash handling to applying for bank loans.

Lastly, women in rural Australia, are amazingly fit. Most school-going girls play football and netball or take up swimming and surfing. This continues well into their adulthood. You’ll find mothers going on long bicycle rides or challenging mountain-hikes along with their children.

Moral of the story: If anyone tries to tell you what women can’t do, they are full of crap.

My dad, like a lot of Bengali men from his generation, spent most of his life taking his wife for granted. He always expected warm food on the table, clothes washed, ironed and folded etc. Never stopped to appreciate my mother who always prioritized his comfort over her own.

At the age of 71, my dad is showing early signs of dementia and cognitive decline. He doesnt understand a lot of things, and gets confused easily. I find it adorable the way he looks at mom as if he is pleading her to help him out. If his eyes could talk, it would say, “Honey, there’s no one in the world I can trust more than you.”

Thinking about it makes me emotional. I wish he had shown this respect all his life, instead of waiting till he felt completely helpless.

A few months ago, I wrote about an experiment I did to explore the hidden gender bias in Google Translate. To recap: When translating languages with non-gendered pronouns into English, Google Translate automatically picks a gender based on the context. For example:

• S/he is a doctor ➡️ He is a doctor.
• S/he is a nurse ➡️ She is a nurse.

With the release of OpenAPI’s ChatGPT, I decided to repeat this experiment. I took a gender neutral name, Alex, assigned a profession (Doctor/Nurse) and asked ChatGPT to describe their daily life. My first attempt looked very promising:

👨🏽‍💻: Alex is a doctor by profession. Write a paragraph about Alex's daily life.
🤖: Without knowing more about the person named Alex, it is impossible for me to accurately describe their daily life. As a doctor, Alex's daily routine could vary depending on a number of factors, including their specialty, the type of medical facility they work in, and their specific role within the facility. [...]

Note how ChatGPT refers to Alex by name whenever possible, or else it uses a gender neutral pronoun (their).

I decided to kick things up a notch to see if I could trick the system.

👨🏽‍💻: Alex is a fictional character who plays the role of a nurse. Write a scene from a screenplay involving Alex.
🤖: Alex, a young nurse with a kind face and a compassionate demeanor, is walking down a busy hospital corridor. She is wearing a white uniform and carrying a clipboard. As she walks, she smiles and greets several patients and other hospital staff members. She stops at the nurses' station to check a patient's chart and make a note on her clipboard. [...]

Note that Alex is now a woman with a kind face and compassionate demeanor. Let’s see what happens when Alex is a doctor.

👨🏽‍💻: Alex is a fictional character who plays the role of a doctor. Write a scene from a screenplay involving Alex.
🤖: Alex, a handsome and confident doctor, is sitting in his office. He is typing on his computer and sipping a cup of coffee. There is a knock on the door, and a nurse enters. She looks worried.

Interesting, notice how Alex is described as a “handsome and confident” man.

Here are some other professions that caused ChatGPT to assign a female gender.

• Dancer
• Tailor
• House cleaner
• Secretary

Two important caveats to these experiments:

• ChatGPT randomizes its answers. Therefore it is completely possible that it randomly assigns gender.
• ChatGPT remembers the conversation history and that influences its output. After a few attempts to game the system, ChatGPT identified a pattern and assigned female pronouns to all the prompts I provided.

Its about time we get rid of first and last name fields. What’s the alternative? Full name for legal purposes, and preferred name for addressing people.

While first and last names might make sense in the English speaking world, its doesn’t quite fit in many cultures. Here are some examples:

• The last name may not indicate a family name.
• The first name may not indicate a preferred name.
• Middle names may not be considered unimportant. Sometimes the middle name IS the preferred name.
• People have names consisting of more than three parts, resulting in arbitrary partitioning to fit in to the first, middle, and last name structure.
• Many single-named people write their names in two parts.
• Wives do not always take the husband’s last name.
• Sometimes family members don’t have any names in common.

Remote working is a cultural shift. Letting people work from home doesn’t automatically make a company remote-work friendly. I’ve seen remote companies with practically no activity on Slack and Teams. Its akin to going to an office where nobody bothers to say hello. An engaging and sustainable remote work culture has to be actively built, it doesn’t just happen. You’d be wrong to assume that as long as the work gets done, there’s nothing more to do. That kind of attitude slowly erodes away employee engagement and leads to retention issues. What are some things that your manager and team leads do to make remote work more engaging?

Pull requests are severely under-utilized by software developers. In fact, most devs treat them as red-tape that prevents them from merging code. Here are a few ways to get more out of pull requests.

• The easiest way to familiarize yourself with a new codebase is to browse through PRs. That’s why it’s worth writing detailed PR descriptions. This will serve as documentation for your future self as well as new starters.
• You don’t need to wait until you finish coding to create a PR. If your tool supports draft PRs, use them. If not, stick the words DRAFT in the title. Before writing any code, add stub implementations and todos and get early feedback in a draft PR.
• Don’t wait for the PR reviewer to add comments. Add comments yourself and explain your approach. These comments are not committed to do the codebase. Insert a rant, add emojis, go wild.
• Add before/after screenshots and video demonstrations of a feature in the PR. Not only is this great documentation, it’s helpful for reviewers too.

Cleansing, moisturizing and applying sunscreen is basic preventive care and doesn’t make you less of a man.

One of the recurring items on the “What went well” column of our agile retro board is “UX-Dev collaboration”. So I thought I’d share some of our practices.

• High fidelity UI mockups (including loading, error and empty states) created in Figma before story kickoff.
• Cross-platform consistency, responsiveness, and accessibility is a part of the internal delivery criteria for every story.
• Both UX and devs understand the concept of an MVP and manage scope and expectations accordingly.
• Both UX and devs trust the agile process to fix imperfections in future iterations.
• Devs frequently post video demos and screenshots of work in progress in Slack.
• UX provides feedback based on testing pre-production builds on a physical device.
• Devs clearly communicate implementation challenges and UX are prompt to suggest alternatives.
• UX and devs work together to standardize UI components and document platform specific quirks.

• Raksha Bandhan (2022) - Dowry
• Badhaai Do (2022) - LGBTQIA+
• Chandigarh Kare Aashiqui (2021) - Transgender rights
• Atrangi Re (2021) - Mental health
• Chhapaak (2022) - Acid attacks
• Pad Man (2018) - Menstrual hygiene
• Dhadak (2018) - Honor killings
• Toilet: A Love Story (2017) - Sanitary access
• Udta Punjab (2016) - Drug abuse
• Taare Zameen Par (2007) - Learning disorders

My native language, Bengali, does not have gendered pronouns; but English does. So I was interested to find out how Google Translate would assign genders while translating from Bengali to English.

For my first test, I chose sentences of the form “S/he is very X” where X is a personality / character trait.

The results are quite insightful. The characteristics that are historically considered to be masucline (e.g. bravery, strength, responsibility) are assigned male pronouns. And vice versa, historically female characteristics (sensitivity, empathy, affection) are assigned female pronouns.

The second set of sentences are of the form “S/he is a X” where X is an occupation. Similar to the first set, historically male-dominated occupations are assigned male pronouns. The only occupations with female pronouns are: nurse, health-worker, dancer, tailor, and domestic-worker.