A few weeks ago I posted a question on Twitter:
The responses flooded in. 399 developers shared their wishes. I categorized every response, and the patterns that emerged tell a clear story about what's broken in modern software development.
The short version:
- Developers can build faster than ever.
- But they can't think uninterrupted, validate quickly, or distribute effectively.
Here's the full breakdown.
The Data
| Rank | Category | Count | % |
|---|---|---|---|
| 1 | Focus / Deep Work / Less Context Switching | 64 | 15.8% |
| 2 | AI & Coding Tool Improvements | 48 | 11.9% |
| 3 | More Time / Work-Life Balance | 33 | 8.2% |
| 4 | Money / Funding / Runway | 30 | 7.4% |
| 5 | Marketing & Distribution | 29 | 7.2% |
| 6 | Feedback & User Validation | 28 | 6.9% |
| 7 | Cofounder / Team / Collaborators | 21 | 5.2% |
| 8 | Energy / Motivation / Mental Health | 20 | 5.0% |
| 9 | Tooling / Infra / DevEx | 16 | 4.0% |
| 10 | Testing / QA / Deployment | 13 | 3.2% |
The top three categories alone account for 36% of all wishes. Let's dig into each one.
1. Focus & Deep Work (15.8%)
The dominant theme. Developers are drowning in tool-hopping, notification pings, and mental model reconstruction.
The quote that says it all:
"If I could wish for one 'not too greedy' thing, it'd be a single workspace that keeps all that state in sync and lets me stay in flow instead of rebuilding my mental model every hour."
Another dev put it this way:
"I wish I could stop spinning on 10 half-built ideas and get help committing deeply to one, so my day-to-day feels focused instead of like a constant tab explosion."
The problem: Modern development requires 10+ tools. Slack, Linear, GitHub, Figma, Sentry, your IDE, documentation, browser DevTools. Every context switch costs 15-25 minutes of recovery time.
A potential solution: Time-block ruthlessly. Declare 2-3 hour "maker windows" as sacred. Turn off notifications. Batch communication into scheduled windows.
SaaS idea: A unified command center that aggregates notifications from all dev tools into a single priority-ranked queue, with AI-powered "is this actually urgent?" filtering.
2. AI & Coding Tool Improvements (11.9%)
This surprised me. Despite the AI revolution, developers are frustrated with current tools — not excited.
The quote that says it all:
"I wished Claude would remember what we discussed during a project like a human coworker. This Groundhog Day experience drives me up the wall."
Developers repeatedly mentioned: AI forgets context, hallucinates, can't handle large codebases, and YOLO mode doesn't actually YOLO.
The problem: Interactive AI coding tools (Cursor, Claude Code) require constant supervision. You're babysitting the AI instead of doing higher-level work.
A potential solution: Shift from interactive to asynchronous AI workflows. Assign tasks and review outputs later, rather than pair-programming in real-time.
SaaS idea: An "AI memory layer" that persists context across sessions, projects, and tools — so your AI assistant actually remembers your codebase conventions, past decisions, and preferences.
Tembo feature idea: Cross-session codebase memory — Tembo already analyzes your codebase before making changes. Extend this to remember architectural decisions and past PR feedback across sessions, so repeated corrections become unnecessary.
Memory shouldn't be optional, it should be default.
3. More Time / Work-Life Balance (8.2%)
Parents especially feel this one. The "side project" dream collides with reality.
The quote that says it all:
"Would love to be able to build my own projects full time. I code for the love of it, but if I could do this then it would be a massive bonus."
And from a parent:
"Just regular time to code every day. It's hard with kids."
The problem: Building software competes with day jobs, family, health, and sanity. There aren't more hours in the day.
A potential solution: Reduce the scope of what you're building. Smaller MVPs, more aggressive feature cuts, ruthless prioritization.
SaaS idea: A "dev time optimizer" that analyzes your calendar, energy patterns, and task complexity to suggest optimal scheduling for different types of work.
4. Money / Funding / Runway (7.4%)
Straightforward: people need financial freedom to build.
The quote that says it all:
"Working on a side project as a student, money is a big barrier. A lot of times I feel like quitting because I couldn't go with the 'best approach' due to limited funds."
The problem: Building takes time. Time requires runway. Many developers are building nights and weekends while working full-time jobs to survive.
A potential solution: Pre-sell before you build. Validate demand and collect payments before writing code.
5. Marketing & Distribution (7.2%)
Engineers consistently feel lost when it comes to getting users.
The quote that says it all:
"That distribution was as deterministic as code. Building the product is 20% of the work; shouting about it effectively is the other 80%."
Another dev wished for "a more introvert-friendly marketing and distribution experience."
The problem: Technical founders can build anything, but reaching customers is a completely different skill. Marketing feels random and uncontrollable.
A potential solution: Treat marketing like engineering. Run experiments, measure results, iterate systematically. Read "Traction" by Gabriel Weinberg.
Tembo feature idea: Auto-generated changelog and release notes — Tembo can already write documentation. Have it automatically generate user-facing release notes, changelog entries, and even Twitter thread drafts when PRs are merged, reducing the marketing burden on technical teams.
6. Feedback & User Validation (6.9%)
Getting honest feedback is surprisingly hard.
The quote that says it all:
"Being able to test ideas on real users and get brutally honest feedback quickly — shortening that feedback loop would be a huge win."
Another dev noted:
"Praise tends to be much more abundant than criticism, but the latter is 10x as valuable."
The problem: Friends and family are too nice. Cold users are too busy. The feedback loop between "shipped" and "validated" is painfully slow.
A potential solution: Build a "Technical Advisory Board" of 15-20 people who commit to monthly feedback sessions. Unpaid, but reciprocal — you test their stuff too.
SaaS idea: A "brutal feedback marketplace" where founders pay real users for recorded, honest usability sessions with specific critique requirements.
Tembo feature idea: User session monitoring integration — Connect Tembo to session replay tools. When users encounter errors or friction, Tembo automatically creates issues and generates ideas to fix, closing the loop between user behavior and code changes.
7. Cofounder / Team / Collaborators (5.2%)
The loneliness of solo building is real.
The quote that says it all:
"Cofounder with the same ambition and risk appetite as me."
Another dev wanted simply:
"I wish I could just get someone with the same drive as me — we're not in it for the money but for the love of the game."
The problem: Great cofounders are rare. Finding someone with complementary skills, shared vision, and matching commitment levels is like finding a needle in a haystack.
A potential solution: Build in public. The best cofounders find you when they see your work and values in action.
8. Energy / Motivation / Mental Health (5.0%)
Burnout lurks beneath the surface of many responses.
The quote that says it all:
"I'd incorporate a personalized AGI network to help productivity in ADHD and engineering. My focus is wack and too many ideas and inventions to keep track of outside my thoughts without assistance."
Another dev admitted to
"popcorn brain — it's hard suddenly not wanting the same thing today that I wanted yesterday."
The problem: Building is mentally exhausting. The constant context switching, uncertainty, and cognitive load depletes energy faster than physical labor.
A potential solution: Treat energy like a finite resource. Schedule demanding work for peak energy hours. Build in recovery time. Exercise helps more than most productivity hacks.
Tembo feature idea: Cognitive load reduction — Tembo handles the "boring but necessary" maintenance work (dependency updates, Sentry errors, documentation) that drains mental energy, preserving developer bandwidth for creative problem-solving.
9. Tooling / Infra / DevEx (4.0%)
Infrastructure work is soul-crushing.
The quote that says it all:
"Infra agents — it's a slog, not fun at all, and takes a lot of time."
The problem: Every hour spent on build systems, deployment pipelines, and infrastructure is an hour not spent on product. But you can't avoid it.
A potential solution: Use managed services aggressively. The cost of Vercel/Railway/Render is worth the time saved versus self-hosting.
SaaS idea: A "legacy code modernization" service that automatically migrates old codebases to modern stacks while preserving functionality.
Tembo feature idea: Infrastructure-as-Code maintenance — Connect Tembo to your Terraform/CDK repositories. When cloud provider APIs change or security patches are needed, Tembo automatically updates IaC and opens PRs.
10. Testing / QA / Deployment (3.2%)
The gap between "works on my machine" and "works in production" remains painful.
The quote that says it all:
"Better e2e testing — tired of sending screenshots and explaining bugs, want the agent to do so itself."
Another dev wanted:
"A QA bot that understood the nuances of UX, copy, and human interaction, and just spotted issues and submitted them automatically to get them fixed."
The problem: Writing tests is tedious. Running them is slow. The feedback loop between code change and confidence is too long.
A potential solution: Prioritize integration tests over unit tests. Test user flows, not implementation details. Use visual regression testing for UI.
SaaS idea: An AI-powered QA agent that plays through your app like a user, identifies UX issues, and files bugs with reproduction steps automatically.
Tembo feature idea: Test generation from PR diffs — When Tembo opens a PR, it automatically generates tests for the changed code paths. When you merge PRs, Tembo can add tests to increase coverage over time, chipping away at test debt without focused effort.
The Meta-Pattern
Looking at these 399 wishes, one theme emerges: developers want to focus on what matters.
The top categories aren't about wanting better syntax highlighting or faster compile times. They're about wanting:
- Uninterrupted thinking time (focus)
- AI that works autonomously (AI tools)
- Time to actually build (work-life balance)
- Validation that their work matters (feedback)
- Help with everything that isn't building (marketing, testing, infra)
The common thread is cognitive bandwidth. Developers have more tools than ever, but less mental space than ever.
What This Means for Tembo
At Tembo, we're building background agents specifically because of these patterns. The shift from interactive AI coding (where you babysit the AI in real-time) to autonomous AI coding (where you assign tasks and review PRs later) directly addresses the top wish: focus.
When your AI coding agent works in the background — fixing Sentry errors at 3am, implementing Linear tickets while you're in meetings, updating documentation when code changes — you get your cognitive bandwidth back.
The future isn't AI that needs constant supervision. It's AI that works like a reliable teammate: you assign the work, it gets done, you review and merge.
That's what we're building.
What would your one wish be? Let me know on X.