Alex Lawsen on avoiding 10 mistakes people make when pursuing a high impact career

Core Premise and Context

• Episode Focus: The conversation features 80,000 Hours advising manager Alex Lawson and host Luis Rodriguez discussing ten common career mistakes made by people attempting to maximize their positive impact.
• Current Status of Speaker: Alex Lawson left 80,000 Hours to join the AI governance team at Open Philanthropy shortly after the recording.
• Framing of "Mistakes": Lawson clarifies that while these are labeled "mistakes," they are often correctable patterns of thought rather than moral failures or signs of personal inadequacy.
• Motivation: The list was compiled from observing common patterns among advisees who share specific characteristics, aiming to provide corrective advice to a broader audience beyond one-on-one consultations.

Mistake 1: Blindly Following "Best Career" Lists

• The Error: Individuals assume they must do the career ranked #1 (e.g., technical AI safety research) because it is objectively the most important, ignoring personal fit.
• The Reality of Fit: Success in high-impact roles often depends on exceptional motivation and interest, not just raw ability; a "good enough" fit in a slightly less urgent field can outperform a "perfect fit" in the most urgent field due to burnout or lack of sustainability.
• The "Good Enough" Fallacy: Being unable to do the top-ranked job due to poor fit does not imply one is "not good"; it simply means they should pursue other roles within that sector (e.g., engineering vs. philosophy in AI) or entirely different sectors.
• Talent Distribution: The world does not operate on a model where the "best" people are forced into the "best" jobs first; therefore, if an individual is a terrible fit for the #1 role, they should not assume they are wasting their talent.
• Signal Detection: People should seek signals that they are "crushing it" (enjoyment, speed of progress) rather than just "going fine," as ambiguous results allow people to remain in suboptimal roles without knowing they could be exceptional elsewhere.
• Objective Metrics: Individuals should look for external validation from colleagues and managers, not just internal narratives, to determine if they are truly excelling.

Mistake 2: Not Trying Hard Enough to Fail

• Definition: The mistake is leaving a "reserve" of effort or avoiding the full intensity of a test to prevent the emotional pain of failure, thereby keeping ambiguity alive.
• The Cost of Ambiguity: By not testing a role intensively, individuals maintain the "nice feeling" that they might be crushing it, avoiding the painful but necessary realization that they are not suited for it.
• Example of Failure Avoidance: Lawson recounts his own experience where he delayed publishing research or hiding early drafts to avoid negative feedback, preventing him from getting the signal that research was not his best path.
• Testing Strategy: Effective testing requires setting clear success/failure criteria (e.g., "I will write a 10-hour report over a weekend") and committing to the outcome, even if it is negative.
• Job Application Tactics: Rejection from hyper-competitive roles is often a signal of "trying hard enough," whereas getting stuck in early rounds of few applications suggests the individual was not applying ambitiously or frequently enough.
• Rewarding Failure: Strategies to make failure less painful include having trusted allies explicitly praise the attempt even in failure, or using "rejection punch cards" to gamify the process.
• Success Criteria: A valid application strategy should result in a distribution of outcomes (e.g., many rejections at the first stage, fewer at later stages) that proves the individual is pushing their limits, rather than consistently failing at the first hurdle due to low effort.

Mistake 3: Optimizing for Immediate Impact Over Long-Term Career Capital

• The Trade-off: Individuals often sacrifice long-term capacity to do good in order to have an immediate, legible impact (e.g., taking a high-status EA internship now vs. a prestigious non-EA role for career capital).
• Career Capital Value: Building skills, gaining signaling experience, and earning money early can pay off significantly more in the long run than direct, immediate impact work.
• Social Pressure: A key driver of this mistake is the desire for peer approval within the effective altruism community, leading people to choose "legibly high impact" roles over "normal" career-building paths that actually maximize long-term utility.
• Specific Example: A university student might spend years running an EA student group (which they dislike) instead of focusing on their technical studies, under the false belief that the group is more valuable to the cause than their individual research breakthroughs later.
• Heuristic for Early Career: In the first few years, the optimal strategy is often to follow normal career heuristics (e.g., high grades, prestigious internships) that maximize learning and skill acquisition.
• Mentorship Weight: Priority should be placed on roles with exceptional mentorship, even if the immediate impact is lower, because rapid skill development accelerates future impact potential.
• Reframing Values: "Living by your values" should be interpreted as doing the thing that maximizes impact over a lifetime, not necessarily the thing that looks most altruistic in the immediate term.

Mistake 4: Misinterpreting Short AI Timelines

• The Assumption: Some believe that if AI will arrive soon (e.g., within 5-10 years), they must stop building career capital and work on AI immediately, regardless of their current skill level.
• The Counter-Argument: If an individual is not yet skilled enough to contribute meaningfully in the short term, attempting to do so may result in near-zero impact or even harm; it is often better to spend those years training for the medium/long term.
• Probability Weighting: Even with a 50% chance of AI arriving in 10 years, if an individual will be significantly more effective in 15+ years, the expected value of training now may exceed the expected value of low-impact immediate work.
• Research Duration: Some problems (like alignment) may require 7-year research projects; starting now means impact is only realized after the likely arrival date, making training necessary regardless of the timeline.
• The "Set Up" Strategy: Spending 7-8 years preparing for the final 2-3 years of high-impact work is a valid strategy if the individual believes the field will become less neglected and they will be a key multiplier in steering others later.
• Concrete Example: A person with a PhD offer in a relevant field should take it even if the research isn't directly on alignment, provided it prepares them for a future research scientist role at a major lab.

Mistake 5: Overvaluing Replaceability

• The Logic: Individuals avoid applying for high-impact jobs because they believe they will be replaced by someone "good enough," meaning their marginal contribution is negligible.
• The Error: This ignores that the difference between the top candidate and the second-best can be massive in high-stakes roles, and that the person they "replace" is likely also highly motivated to do good elsewhere.
• Talent Allocation: The most efficient way to allocate talent is for everyone to apply to the jobs they are best suited for, rather than a self-imposed ban on applying; this allows hiring managers to select the absolute best fit.
• Conditioning on Values: If the other candidates in the pool also care about doing good, the individual should assume that if they don't take the job, someone else who also cares will, and the net effect is simply that the best person gets the job.
• Exception for Multiple Offers: The only scenario where replaceability is a critical factor is when an individual holds multiple offers for similar roles; in this case, they should ask hiring managers about the strength of the remaining talent pool to decide which role is more dependent on them.

Mistake 6: Constantly Switching Paths Due to "Shiny New Options"

• The Pattern: Individuals on high-impact paths constantly consider switching to even more impactful roles, preventing them from settling and mastering a specific skill set.
• The Cost of Switching: Frequent job switching (e.g., every 9-12 months) leads to a lack of deep expertise, emotional instability, and an inability to "get in the groove" of high-performance work.
• Moral Obligation Guilt: Many feel guilty for not seriously considering every new high-impact option, viewing it as a failure to maximize good, which paralyzes them from fully committing to their current path.
• The "Rowing" Metaphor: Trying to constantly check the map (evaluate options) while rowing leads to slow progress; the optimal strategy is to row hard in one direction for a set period, then stop to check the map.
• Policy Recommendation: Establish a fixed re-evaluation period (e.g., "I will stay in this role for two years unless a compelling case arises") to prevent distraction and ensure deep skill acquisition.
• Information Gathering Trade-off: The cost of gathering more information about future options is real; spending excessive time researching before acting often yields diminishing returns compared to the value of doing the work and learning from it.

Mistake 7: Deferring Cause Prioritization

• The Mistake: Individuals refuse to do their own cause prioritization because they believe "smart people at 80k/Open Philanthropy" have already solved it and they should just follow that advice.
• The Correction: It is acceptable and necessary to think through cause prioritization for oneself, even if it does not result in a complete philosophical solution; the goal is to align with one's intuitions and avoid internal tension.
• Self-Reliance: Relying solely on external advice without personal engagement leads to a lack of ownership and can result in pursuing paths that feel misaligned.

Mistake 8: Ignoring Conventional Career Wisdom

• The Bias: Effective altruists often ignore conventional career advice (e.g., getting mentorship, working for established firms) because they view it as "normal" or "selfish."
• Mentorship Value: Conventional wisdom correctly identifies that good mentorship accelerates learning; ignoring this in favor of a specific cause can slow down long-term impact potential.
• Established Institutions: Working at well-run, established companies (e.g., Jane Street) can provide superior training and thinking frameworks compared to scrappy startups, even if the latter feels more "altruistic."
• Suspicious Convergence: If multiple self-interested reasons and altruistic reasons point to the same choice, it does not necessarily mean the altruistic reason is a rationalization; often, self-interest and impact are aligned.
• Testing Hypothesis: A useful heuristic is to ask, "What would I choose if I knew my choice would automatically be the best for the world?" to identify conventional wisdom that should be considered.

Mistake 9: Pursuing Community Building Despite Lack of Fit

• The Reasoning: Individuals justify doing community building for AI because they don't feel fit for technical research, believing they can still create significant impact by convincing others to work on it.
• The Error: If the individual hates community building or gains no transferable skills from it, the role is a net negative; it sacrifices immediate impact and long-term skill growth.
• Skill Transferability: If the skills gained from community building are not expected to be used in a future high-impact role, the individual should prioritize setting themselves up for that future role instead.
• Decision Metric: The viability of community building depends on whether the individual enjoys it, has spare time, and is building skills they will use later; if not, it is a mistake.

Meta-Themes and General Observations

• Over-Correction to Extremes: A recurring pattern is noticing that conventional wisdom is flawed, updating in the right direction, but then updating too far to the opposite extreme.
• Vocabulary Gap: The community often lacks the vocabulary to express nuanced positions (e.g., "I agree with you but less strongly than you do"), leading to binary debates and exaggerated updates.
• Correlated Information: Individuals often treat multiple sources of advice from the same community (e.g., multiple 80,000 Hours podcast episodes) as independent data points, leading to over-confidence in extreme updates.
• Moral Perfectionism: Many mistakes stem from a hyper-focus on moral perfectionism, where individuals feel they must be absolutely certain of doing the "best" thing or that any compromise is a failure of character.
• Uncertainty Management: Career decisions in high-stakes areas (like AI) involve unusually high uncertainty; acting with "reasonable certainty" is often the best strategy, even if it doesn't feel "good."
• Actionable Advice: The consensus is that individuals should gather a reasonable amount of information, make a decision, commit for a set period, and then re-evaluate, rather than staying in a state of perpetual analysis.