1
The most common mistake new AI builders make about AI / LLM work — and the specific fix.
contrarianSingle tweet
Idea bank · AI
30 AI tweet ideas
Copy-paste AI tweet ideas with hook + format hints. For AI builders, prompt engineers, and technical readers building on LLMs.
Weekly tips
Weekly X (Twitter) growth playbooks
One specific tactic each Sunday — pulled from accounts actively growing on X. No fluff, no resends, unsubscribe anytime.
We'll never share your email. Unsubscribe with one click.
AI content on X is bimodal: builders + general-curiosity readers. Builder-targeted content (specific model decisions, prompt techniques, eval methodologies) wins with engineers; general-curiosity content gets superficial engagement. These 30 ideas tilt toward builder-targeted — higher signal, higher conversion to product audiences.
30 tweet ideas
2
An exact number from a AI / LLM work experience this week — and the lesson behind it.
specific numberSingle tweet
3
A 5-step framework for solving the biggest AI / LLM work problem you've faced. One step per tweet in a short thread.
frameworkThread (5-9)
4
Why most AI builders are wrong about a specific aspect of AI / LLM work. Defend with specifics.
contrarianSingle tweet
5
A specific tool / process / habit that 10x'd your AI / LLM work results. Name the tool, show the specifics.
specific numberSingle tweet
6
The hardest decision you made about AI / LLM work in the past year. What you chose + why + how it turned out.
personal stakeThread (5-9)
7
An open question about AI / LLM work you don't have a great answer for. Lean into the uncertainty publicly.
curiosity gapSingle tweet
8
The 3 books / podcasts / courses that shaped how you think about AI / LLM work. Why each matters.
frameworkThread (5-9)
9
A specific failure in AI / LLM work that taught you more than any success. Detailed retrospective.
personal stakeThread (7-12)
10
The contrarian belief you hold about AI / LLM work that most peers disagree with — and the evidence behind it.
contrarianSingle tweet
11
A behind-the-scenes look at how you actually work on AI / LLM work. Show the workflow, not the highlights.
curiosity gapThread (5-9)
12
An ROI calculation showing the dollar impact of a specific AI / LLM work decision. Show the math.
specific numberSingle tweet
13
A specific question to ask before investing time/money in AI / LLM work. The question most AI builders skip.
frameworkSingle tweet
14
Why a popular AI / LLM work approach you used to follow no longer works. What you do instead.
contrarianThread (5-9)
15
The metric you obsess over in AI / LLM work that nobody else watches. Why it matters.
personal stakeSingle tweet
16
An anonymous case study: someone you know who got AI / LLM work right (or wrong). The transferable lesson.
curiosity gapThread (5-9)
17
The earliest signal that something is going wrong with AI / LLM work — before the obvious metrics turn red.
specific numberSingle tweet
18
A 2-line framework for making faster AI / LLM work decisions when stuck. What to ask, what to skip.
frameworkSingle tweet
19
Why AI / LLM work expertise compounds — and the specific habits that build that compounding.
personal stakeThread (5-9)
20
The first sign you've outgrown the standard AI / LLM work playbook. What changes when you have.
curiosity gapSingle tweet
21
The single best piece of AI / LLM work advice you ever received — and the worst.
contrarianSingle tweet
22
A common AI / LLM work myth, debunked with a specific counter-example you've personally seen.
contrarianSingle tweet
23
Three patterns that consistently predict success in AI / LLM work. The pattern, the example, the why.
frameworkThread (5-9)
24
A specific number that defines what 'good' looks like in AI / LLM work. The number, the source, the context.
specific numberSingle tweet
25
What AI / LLM work would look like if you started over today knowing what you know now.
personal stakeThread (7-12)
26
An emerging trend in AI / LLM work that AI builders are sleeping on. The data + the implication.
curiosity gapThread (5-9)
27
The hardest question AI builders face about AI / LLM work — and how to answer it for yourself.
frameworkSingle tweet
28
A controversial-but-defensible take on the future of AI / LLM work. Lead with conviction.
contrarianSingle tweet
29
A specific AI / LLM work habit you started 12 months ago that's compounded. The habit, the time, the result.
personal stakeThread (5-9)
30
What you wish someone had told you about AI / LLM work on day one. Direct, specific, no platitudes.
curiosity gapThread (5-9)
Common questions
Should I share my AI prompts publicly?+
Most of the time, yes. Prompts that work are rarely competitive moats — the actual moat is the data + evals + infrastructure around them. Sharing prompts demonstrates expertise and earns builders' trust. Exceptions: highly tuned production prompts that took weeks to develop.
Is the AI niche oversaturated?+
AI commentary is saturated. AI building is not. The builder audience on X is still very narrow — most accounts posting about AI haven't shipped production AI. If you have, your specifics break through the noise immediately.
How technical should AI content be?+
As technical as you can authentically be without losing the audience you want. Engineers + technical founders are the high-converting AI audience; the bar is engineer-readable, not engineer-perfect. Specific model names, specific eval methods, specific cost / latency numbers — these are the differentiated content.