6 days ago
Streaks reset to zero and I quit Streaks app gave up on me when I missed Tuesday. Reset to 0 from 23 days. The rigidity is the bug, not me. There has to be a better model than "miss once, restart."
6 days ago · May 13, 2026
Streaks reset to zero and I quit Streaks app gave up on me when I missed Tuesday. Reset to 0 from 23 days. The rigidity is the bug, not me. There has to be a better model than "miss once, restart."
5 days ago
What if habit trackers measured WHY? Most apps track frequency. What about asking "did this feel good?" after each one. Sentiment over streaks. The streak isn't the goal — the behavior change is, and behavior follows feeling.
5 days ago · May 14, 2026
What if habit trackers measured WHY? Most apps track frequency. What about asking "did this feel good?" after each one. Sentiment over streaks. The streak isn't the goal — the behavior change is, and behavior follows feeling.
4 days ago
Notion is not a habit tracker Spent 3 hours setting up the Notion habit tracker template. Used it for 4 days then stopped. Tool overhead > behavior change. The setup itself was the failure mode.
4 days ago · May 15, 2026
Notion is not a habit tracker Spent 3 hours setting up the Notion habit tracker template. Used it for 4 days then stopped. Tool overhead > behavior change. The setup itself was the failure mode.
3 days ago
Habit recovery — 5 of 7 is still a streak Missed days shouldn't penalize. Build in "recovery" — if you do it 5 of 7 days, that's a streak. Grace period. Most weeks aren't perfect.
3 days ago · May 16, 2026
Habit recovery — 5 of 7 is still a streak Missed days shouldn't penalize. Build in "recovery" — if you do it 5 of 7 days, that's a streak. Grace period. Most weeks aren't perfect.
1 day ago
App I want: habit tracker that asks how you felt Daily check-in. Was the run hard or easy. Was the meditation focused or scattered. Track the experience, not just the action. Quick scale: ✗ / okay / good / great. Three taps max.
1 day ago · May 18, 2026
App I want: habit tracker that asks how you felt Daily check-in. Was the run hard or easy. Was the meditation focused or scattered. Track the experience, not just the action. Quick scale: ✗ / okay / good / great. Three taps max.
7 days ago
Claude code scope creep Asked Claude to fix a bug. It also refactored 3 unrelated files. Need to be explicit about scope in prompts. "Touch only the file I named" works better than "fix this bug."
7 days ago · May 12, 2026
Claude code scope creep Asked Claude to fix a bug. It also refactored 3 unrelated files. Need to be explicit about scope in prompts. "Touch only the file I named" works better than "fix this bug."
4 days ago
AI for code review surfaces fast issues Used GPT-4 for PR review. Caught 3 nil-checks I missed. Missed the architectural issue with the cache invalidation. Good complement, not replacement.
4 days ago · May 15, 2026
AI for code review surfaces fast issues Used GPT-4 for PR review. Caught 3 nil-checks I missed. Missed the architectural issue with the cache invalidation. Good complement, not replacement.
2 days ago
AI judgement vs translation AI is great at translation (this lang → that, this format → that). Bad at judgement (should we even do this?). Use it where there's a clear target. Skip it where the target itself is the question.
2 days ago · May 17, 2026
AI judgement vs translation AI is great at translation (this lang → that, this format → that). Bad at judgement (should we even do this?). Use it where there's a clear target. Skip it where the target itself is the question.
5 days ago
Article: Why habits fail (2-day rule) Read James Clear piece — the 2-day rule. Never miss twice in a row. Strong rule, simple to implement. Could be a core feature, not just advice.
5 days ago · May 14, 2026
Article: Why habits fail (2-day rule) Read James Clear piece — the 2-day rule. Never miss twice in a row. Strong rule, simple to implement. Could be a core feature, not just advice.
3 days ago
WebAuthn for app auth Looked into WebAuthn / passkeys this morning. iOS support is solid. Passwords are dead — worth investigating for any new mobile app. Reduces account-recovery surface area.
3 days ago · May 16, 2026
WebAuthn for app auth Looked into WebAuthn / passkeys this morning. iOS support is solid. Passwords are dead — worth investigating for any new mobile app. Reduces account-recovery surface area.
1 day ago
Tailwind v4 weirdness Set up TW v4 — config-in-CSS feels backwards but the @theme block is actually clean. Saves a config file. Test it on next project.
1 day ago · May 18, 2026
Tailwind v4 weirdness Set up TW v4 — config-in-CSS feels backwards but the @theme block is actually clean. Saves a config file. Test it on next project.
12 hours ago
Realm vs Firebase for local storage For the habit tracker idea — local-only with Realm vs Firebase with sync? Local feels right for a single-user habit app. But losing data when the phone dies is a real fear. Need to think about this.
12 hours ago · May 19, 2026
Realm vs Firebase for local storage For the habit tracker idea — local-only with Realm vs Firebase with sync? Local feels right for a single-user habit app. But losing data when the phone dies is a real fear. Need to think about this.
7 days ago
Why I can't get into flow in the mornings Spent 30 min trying to focus on a problem this morning. Keep checking Slack before I'm actually awake. The distraction isn't the phone — it's the habit of reaching for it. Need a buffer routine.
7 days ago · May 12, 2026
Why I can't get into flow in the mornings Spent 30 min trying to focus on a problem this morning. Keep checking Slack before I'm actually awake. The distraction isn't the phone — it's the habit of reaching for it. Need a buffer routine.
6 days ago
Context window as the real bottleneck The model isn't the constraint anymore. The context window is. When the AI forgets what you built two hours ago, you spend more time re-explaining than building. This is the actual problem to solve.
6 days ago · May 13, 2026
Context window as the real bottleneck The model isn't the constraint anymore. The context window is. When the AI forgets what you built two hours ago, you spend more time re-explaining than building. This is the actual problem to solve.
5 days ago
Tailwind: missing component for card rotations Can't do arbitrary rotation values in Tailwind without config or inline styles. -rotate-1 and rotate-2 aren't in default scale. Just use style= for one-off transforms.
5 days ago · May 14, 2026
Tailwind: missing component for card rotations Can't do arbitrary rotation values in Tailwind without config or inline styles. -rotate-1 and rotate-2 aren't in default scale. Just use style= for one-off transforms.
5 days ago
Sleep tracking app idea — passive, not opt-in Every sleep tracker I've tried requires you to tell it when you go to bed. That's the first thing to automate. Accelerometer + mic data should be enough. No tap required.
5 days ago · May 14, 2026
Sleep tracking app idea — passive, not opt-in Every sleep tracker I've tried requires you to tell it when you go to bed. That's the first thing to automate. Accelerometer + mic data should be enough. No tap required.
4 days ago
Reading: system design interview book — Chapter 4 Rate limiter design. Token bucket vs sliding window. Didn't realize how much the choice depends on burst tolerance. Keep for next system design conversation.
4 days ago · May 15, 2026
Reading: system design interview book — Chapter 4 Rate limiter design. Token bucket vs sliding window. Didn't realize how much the choice depends on burst tolerance. Keep for next system design conversation.
3 days ago
Prompt pattern: "here's what I already tried" Giving AI your failed attempts before asking for help cuts the back-and-forth in half. It stops suggesting the obvious. Also works in Stack Overflow posts. Should be a reflex.
3 days ago · May 16, 2026
Prompt pattern: "here's what I already tried" Giving AI your failed attempts before asking for help cuts the back-and-forth in half. It stops suggesting the obvious. Also works in Stack Overflow posts. Should be a reflex.
3 days ago
Offline-first is underrated for single-user apps If there's no collaboration requirement, offline-first is almost always better. Faster, cheaper to run, works in dead zones. Habit tracker doesn't need a server for phase 1.
3 days ago · May 16, 2026
Offline-first is underrated for single-user apps If there's no collaboration requirement, offline-first is almost always better. Faster, cheaper to run, works in dead zones. Habit tracker doesn't need a server for phase 1.
2 days ago
Question: does anyone actually use the Pomodoro technique? I've seen it everywhere but I've never met a developer who swears by it long-term. Maybe it's a starter crutch? Or maybe I'm the wrong sample. Curious.
2 days ago · May 17, 2026
Question: does anyone actually use the Pomodoro technique? I've seen it everywhere but I've never met a developer who swears by it long-term. Maybe it's a starter crutch? Or maybe I'm the wrong sample. Curious.
2 days ago
Cloudflare Workers cold start is fast enough Been worried about cold start latency for the API layer. Under 10ms in practice. Not a concern for the use case. Stop over-engineering the infra and ship the feature.
2 days ago · May 17, 2026
Cloudflare Workers cold start is fast enough Been worried about cold start latency for the API layer. Under 10ms in practice. Not a concern for the use case. Stop over-engineering the infra and ship the feature.
1 day ago
Feature idea: "why did Bridge suggest this?" If an AI surfaces a project proposal, the user should be able to ask why. Which notes triggered it, what pattern it detected. Transparency builds trust faster than accuracy.
1 day ago · May 18, 2026
Feature idea: "why did Bridge suggest this?" If an AI surfaces a project proposal, the user should be able to ask why. Which notes triggered it, what pattern it detected. Transparency builds trust faster than accuracy.
1 day ago
Book quote — Thinking, Fast and Slow "The confidence people have in their beliefs is not a measure of the quality of evidence." Good reminder when reviewing AI output. High confidence ≠ correct.
1 day ago · May 18, 2026
Book quote — Thinking, Fast and Slow "The confidence people have in their beliefs is not a measure of the quality of evidence." Good reminder when reviewing AI output. High confidence ≠ correct.
8 hours ago
TypeScript: prefer unknown over any at API boundaries Caught a bug because I used any on an API response type. unknown forces you to narrow before use. Stricter but the compiler actually helps you. Worth the extra lines.
8 hours ago · May 19, 2026
TypeScript: prefer unknown over any at API boundaries Caught a bug because I used any on an API response type. unknown forces you to narrow before use. Stricter but the compiler actually helps you. Worth the extra lines.
5 hours ago
Future-self question: did the habit tracker ever ship? If you're reading this and it didn't — figure out what stopped you. Was it scope, or distraction, or the same thing that stopped the last three projects?
5 hours ago · May 19, 2026
Future-self question: did the habit tracker ever ship? If you're reading this and it didn't — figure out what stopped you. Was it scope, or distraction, or the same thing that stopped the last three projects?
