Ana De Canha | UX Engineer

7/

Be careful with the activity.

Does it drain you? Then do it just a little every day (like 30 min max)

The rest, you can batch as usual 🙌🏼

6/

✅ My main learning: if you want to stay productive, you DON’T have to batch everything

5/

I do like (even love) everything else, but the last bit of it — video editing — was very draining

And when I batched tasks, I batched that too, which left me completely burned out by the end of the cycle

4/

Here I found my answer: editing videos drains my energy

3/

I was trying to discover what the problem was 🤔

Overall, I liked the activity I was doing (content creation). Then why did I suddenly stop after just a few weeks?

2/

I know batching repeating tasks is the best way to get more things done, and I've always been doing that

But sometimes I just stop doing everything for a week or so. It’s like I suddenly burn out 😶‍🌫️

1/

I realised there’s no need to always batch tasks to be more productive

There’s a rule for that I didn’t know about → If that task drains your energy, it’s not worth to batch 👀

I read something this week that challenged my thinking. I want to know what you think:

Even if people’s fears are irrational, they still feel very real and painful

Should experts protect us only from real risks, or also from the fear itself? What’s your take?

People usually focus on the idea and not the problem. I’m the opposite 🙃

I know the problem really well, but I’m struggling to sell the idea

Am I the worst coworker for forgetting names? 🥲

Me: "there's this thing that does X, X, and X. We should use it"

Coworker: "Perfect! What’s it called?"

Me: "No idea… but it does this and works like this. Just find it"

Teleportation is the best superpower, and you can’t convince me otherwise 👀

6. Refine & iterate 🔄

Combine data + insights → update hypotheses, product design, or messaging

Then repeat the cycle with fresh data

Which step do you find hardest: analyzing data or talking to users? 👀

5. Ask directly (qualitative research) 🧐

Talk to users:

👂 User interviews

👀 Contextual inquiry

✍️ Open-ended survey questions

Numbers ≠ motivations

4. Design and run tests 🧪

Validate your assumptions through:

✅ A/B testing

✅ Prototype tests

✅ Usability sessions

✅ Think-alouds

3. Generate hypotheses 💭

Turn patterns into assumptions about why it happens

Example:

🔹 “Form feels too long”

🔹 “CTA label is unclear”

2. Analyze results 🔍

Look for anomalies, bottlenecks, or unexpected behaviors

Example:

❌ “Users drop at Step 3”

❌ “Avg. time on task is unusually high”

1. Collect quantitative data 📊

Analytics, funnels, heatmaps, event tracking, scaled surveys

🎯 Goal: spot what’s happening → drop-offs, errors, or usage patterns

Data will tell you what users do But it won’t tell you why they do it

Here’s a simple UX framework to move from what → why 🧵👇

This is the copy when there's no lyrics in a Spotify song

Would you do it differently? What would you write? 👀

The AI said I’m doing great and should take a break

And who am I to argue with that? 😌

4️⃣ Provide real context or verified info

Example:
“Based on this 2023 WHO report that says X, explain why…”

This anchors the model to real data and helps it build a better response around that info

Tell what you do when you're not sure you can trust an AI response? 👇🏼

3️⃣ Give permission to say ‘I don’t know’

Say:
– “Only answer if you’re at least 80% confident”
– “If you’re not sure, just say that”

Most models are tuned to sound confident, even when unsure

Telling them it’s ok not to know makes them more honest 💃🏻

2️⃣ Narrow the question

Instead of:
“Tell me about quantum physics”

Try:
“Explain what a photon is, in 2–3 sentences, for a high school student”

Broad = wild guesses
Specific = grounded answers

1️⃣ Ask for sources or evidence

Say:

– “List your sources and make sure they’re real”
– “Cite real publications with dates”

Why it helps: The model cross-checks internally instead of pulling random stuff from memory

🟡 Bonus: You can verify the sources yourself

I just learned 4 tricks to reduce hallucinations when talking to an AI model🤖🫣

They're all super simple and they work because they help the model stick to what it actually knows instead of guessing

Here’s what to try 👇🏼

Most people will only try your product once — especially in beta 😬

If they don’t know what to do right away, they’ll bounce. A simple onboarding can change that

✅ Highlight 1 key feature

👉🏼 Ask for 1 action

💬 Get 10x better feedback

Let’s help each other out 🗣️

1. How do you keep track of your work?
2. What’s your go-to way to present yourself in interviews?

Share your tips so others can take notes 👀

Looking for a job after years of experience… and realizing you never took notes of your projects, progress, or achievements

How do you defend your case then? 🫠

I’ve seen this happen to so many friends (I do have my notes) but it’s a real struggle in interviews

If you’re in tech, does it feel like this for you too?

Also, what about other industries?

I’m curious to hear your experiences 👀

Looking for a job in tech right now feels… awful

I fell in love with this industry because people were kind, curious, and helped each other.
Now it’s like a cruel contest to see who can work the hardest. I really don’t like it 🫠