Last 100 Days of 2026: Here's how to use them wisely

Key strategies for the last ~100 days of 2026 (career + money focus)

1) Start with a clear financial outcome

• Define a specific end-of-year money goal (e.g., $150k/year salary, a paid internship, business revenue).
• Work backwards from that goal: with 100 days, you can meaningfully improve skills or land a better role.

2) Match your plan to your situation

• There’s no one-size-fits-all roadmap.
• Adapt based on constraints like:
  • experience level
  • job status
  • location
  • other personal factors

---

If you’re jobless or in a low-paying job (especially IT)

Build a “role-ready” skill stack

Combine technical skills with:

• Communication skills (explaining work clearly, speaking in calls)
• Consistency and dependability
• Ability to explain your product/codebase in minutes

Use AI tooling to accelerate development, including examples mentioned:

• Claude Code / Codex
• Opus 5.5 (described as useful for programming tasks)

Stop mass, unfocused applications—do targeted outreach instead

• Avoid applying broadly using random volume or auto-application tools.
• Prefer funded startups, especially seed / Series A companies from the last ~6 months, since they’re actively hiring.

Suggested daily practice (example: 10 letters/day)

• Write 10 highly personalized letters per day to founders.
• Do not use AI to produce generic “nonsense”—founders can detect it.
• Use the company’s product first:
  • Try it (free trials if available; buy if it’s reasonable)
  • Identify:
    • 10 problems / friction points
    • 10 benefits
  • Suggest ways to scale/fix flaws within ~3 months
• Positioning idea: the letter should read like you’re already part of the team.

---

If you already have a job (and want growth)

Use a daily learning block (protect consistency)

• Set aside ~2 hours/day (or at least an average of 2).
• Spread learning across the week—not just one weekend day.
• Total target: ~14 hours/week.

Use AI as a “review + debate” partner

Feed your work into AI (example: Claude, including a higher-tier option like Opus 5.5). Ask it to:

• check architecture/design for correctness
• identify improvements or risks
• provide deeper, more nuanced critiques

Use it like a debate to extract rare edge-case knowledge for your next day at work.

Use AI to simulate learning/debugging

• Don’t only use AI when you’re stuck.
• Use it for “dry” reasoning/review:
  • read/understand what you did
  • detect likely failure modes and nuance
• Goal: learn specific fixes that help immediately.

Learn adjacent domains (risk-proof your career)

• Study a different field gradually (examples given: finance, biology, law, etc.).
• Reason: if layoffs or shutdowns happen, cross-domain expertise improves resilience.
• Skill diversification increases competitiveness (e.g., programmer + finance → fintech advantage).

Optional path: turn expertise into consulting/services

• Start by selling time via consulting/services.
• Use AI tools to build solutions and internal systems for clients.
• You still must be qualified and follow any relevant licensing rules.

---

Mindset guidance around AI (how to avoid getting stuck psychologically)

Avoid the extremes

• “AI will do everything, so I don’t need to learn” → leads to stagnation and anxiety, especially if you’re poor or unemployed.

Adopt an optimistic, action-oriented stance

• Use AI to move faster, discover solutions, and create value.
• The speaker emphasizes that money can still be made in the AI ecosystem (e.g., teaching, selling AI productivity help, or offering AI-related services).

---

Sponsored tool mentioned (productivity/testing workflow)

TestRight (agent-based testing infrastructure)

• Lets AI generate and run tests like a real user.
• Can:
  • detect errors
  • write code to fix them
  • create/update tests
  • automate future regression checks
• Purpose stated: prevent buggy AI-assisted code from reaching production.

---

Presenters / Sources

• Presenter: Not explicitly named in the subtitles (speaker referenced only implicitly)
• Sponsor: TestRight
• Other tools/models mentioned: Claude (Anthropic), Opus 5.5, Claude Pro / Pro Max, Claude Code, Codex
• Additional platforms mentioned: TestFlight (mentioned as used by companies; sponsor attribution in subtitles is inconsistent, but TestRight is clearly the sponsored product)