{"course":{"slug":"ai-careers","title":"Working in AI","url":"https://addaly.com/learn/ai-careers","level":"beginner","track":"work"},"lesson":{"slug":"rejection-and-the-log","title":"What a rejection tells you, and what it does not","url":"https://addaly.com/learn/ai-careers/rejection-and-the-log","position":60,"of":85,"minutes":8,"answer":"Most rejected finalists were hireable, because hiring ranks candidates under a budget rather than assessing them absolutely, so read your log in tens, not one at a time.","keyIdea":"Most rejected finalists were hireable, because hiring ranks candidates under a budget rather than assessing them absolutely, so read your log in tens, not one at a time."},"body":{"format":"markdown","text":"## The arithmetic of losing\n\nAt the final stage, a strong candidate converts to an offer perhaps a third of the time. That means two out of every three people who reach a final round are rejected, and most of them were entirely capable of doing the job.\n\nThis is not consolation. It is the mechanism. Hiring is a **ranking against other candidates under a budget**, not an absolute judgement of you. A rejection is compatible with all of the following, none of which is about your ability:\n\n> Figure: A hundred people who reached a final round. Bars (of 100 finalists): Offered 33; Rejected, and capable of the job 52; Rejected for a gap they could name 15. About a third convert. Of the rest, most could have done the job: somebody was slightly better on the day, the internal candidate appeared, the headcount froze, the manager left. Hiring ranks candidates under a budget; it does not assess anyone absolutely.\n\n- Somebody else was slightly better on the day, or knew someone on the panel.\n- The internal candidate applied in week three.\n- The headcount was frozen between your second and third interviews.\n- The hiring manager left and the role was withdrawn.\n- The team decided mid-process that they needed someone more senior.\n\nCompanies rarely explain which of these happened, so candidates supply the explanation themselves, and the one people reach for is always the same and always the least likely.\n\n## Asking for feedback so that you sometimes get it\n\nMake it cheap to answer. A long email asking for a detailed review gets nothing; two lines gets a reply maybe a third of the time.\n\n> \"Thanks for letting me know. If there was one specific gap that decided it, I would find that genuinely useful: a single line is plenty. Either way I appreciated the process.\"\n\nTwo things to know. First, many companies have a policy against detailed feedback, on legal advice, because a specific reason can become evidence in a discrimination claim. Silence is often policy rather than disdain. Second, when you do get feedback, treat it as one data point. A single interviewer's view is noisy; the same comment from three companies is a finding.\n\n## The log\n\nThe tracker from the funnel lesson gets four extra columns for anything that reached a first conversation:\n\n- **Stage reached.**\n- **Their stated reason**, verbatim, or \"none given\".\n- **Your own read**, written the same day, in one sentence.\n- **One thing to change**: and it must be one thing, not a list.\n\nThen review after every ten rows, not after every rejection. Reading it after each one is how a bad week becomes a theory about yourself. Reading ten together shows a pattern or shows that there is not one, and both are useful.\n\nPatterns you will actually see: the same weak answer to the same question; falling out at the same stage every time; only applying when you feel confident, which correlates with roles below your level; four rejections all from companies whose process you disliked, which is a different problem.\n\n## Ghosting, and a rule to end it\n\nYou will be ignored, sometimes after a final interview. It is rude, it is common, and it is structural rather than personal: a recruiter with 1,400 applications and a frozen requisition sends nothing.\n\nHave a rule so it stops occupying you: **two follow-ups, one week apart, then close the row and stop thinking about it.** Write the rule down before you need it, because the decision is harder in the moment.\n\n## Re-applying\n\nApplying again after six to twelve months is completely normal. Their tracking system keeps your history, which cuts both ways: a strong previous process is remembered favourably, and six scattergun applications are also visible.\n\nWhen a rejection says \"please do apply for future roles\", it is sometimes politeness and sometimes literal. You can tell the difference by whether they name a role type or a timeframe.\n\n## The part that is not tactics\n\nLong searches are damaging, and the damage is not proportional to how tough anybody is. Three things reliably help, and none of them is more applications.\n\n**A fixed weekly quota** (the eleven items from the volume lesson) so effort becomes a decision made on Sunday rather than a mood negotiated daily. On the days you do the quota, the day was a success regardless of any reply.\n\n**Other people.** A study partner, a group, someone who is also searching. Isolation makes every rejection larger, and a search run entirely alone is the version that breaks people.\n\n**A separation you state out loud.** The market's response is information about the market. It is not a measurement of your worth, and the two feel identical from the inside after the fifteenth rejection. If it is affecting your sleep, eating or health, that is a signal to talk to somebody (a doctor, a friend, a helpline) and not a signal to send more applications."},"check":{"question":"A candidate has been rejected after three final-round interviews at three companies, with no feedback given. What is the most defensible interpretation?","options":[{"text":"Something specific about their interviewing is failing at the last stage and they should overhaul their approach before applying again.","correct":false,"tempting":"Draws a strong conclusion from three trials of a stage where roughly two thirds of good candidates are rejected anyway."},{"text":"The lack of feedback shows all three companies found the same serious problem and are avoiding saying it.","correct":false,"tempting":"Reads a common legal policy against detailed feedback as concealment of a specific fault."},{"text":"Reaching three final rounds indicates the funnel is working, three losses at roughly a one-in-three conversion is an ordinary outcome, and the useful action is to log each one and look for a pattern once there are ten data points.","correct":true},{"text":"They should apply to less competitive companies where they are more likely to convert.","correct":false,"tempting":"Lowers ambition in response to what is statistically an expected run of results, and gives up the level they have demonstrably reached."}]},"prev":{"slug":"what-the-money-looks-like","title":"Pay, rates, and who actually decides them","url":"https://addaly.com/learn/ai-careers/what-the-money-looks-like"},"next":{"slug":"what-is-actually-in-an-offer","title":"The headline number is the least useful one","url":"https://addaly.com/learn/ai-careers/what-is-actually-in-an-offer"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'What a rejection tells you, and what it does not', in Working in AI. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-careers/rejection-and-the-log (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"What a rejection tells you, and what it does not","courseTitle":"Working in AI","module":{"index":6,"slug":"the-interview","title":"The interview, from first screen to signed offer"},"answer":null,"keyIdea":"Most rejected finalists were hireable, because hiring ranks candidates under a budget rather than assessing them absolutely, so read your log in tens, not one at a time.","hero":{"kind":"figure","figure":{"t":"figure","label":"Figure","figure":{"type":"bars","unit":"of 100 finalists","title":"A hundred people who reached a final round","series":[{"label":"Offered","value":33},{"label":"Rejected, and capable of the job","value":52,"highlight":true},{"label":"Rejected for a gap they could name","value":15}],"caption":"About a third convert. Of the rest, most could have done the job: somebody was slightly better on the day, the internal candidate appeared, the headcount froze, the manager left. Hiring ranks candidates under a budget; it does not assess anyone absolutely."},"alt":"Figure: A hundred people who reached a final round. Bars (of 100 finalists): Offered 33; Rejected, and capable of the job 52; Rejected for a gap they could name 15. About a third convert. Of the rest, most could have done the job: somebody was slightly better on the day, the internal candidate appeared, the headcount froze, the manager left. Hiring ranks candidates under a budget; it does not assess anyone absolutely."}},"blocks":[{"t":"h2","id":"the-arithmetic-of-losing","runs":[{"text":"The arithmetic of losing"}]},{"t":"p","runs":[{"text":"At the final stage, a strong candidate converts to an offer perhaps a third of the time. That means two out of every three people who reach a final round are rejected, and most of them were entirely capable of doing the job."}]},{"t":"p","runs":[{"text":"This is not consolation. It is the mechanism. Hiring is a "},{"text":"ranking against other candidates under a budget","b":true},{"text":", not an absolute judgement of you. A rejection is compatible with all of the following, none of which is about your ability:"}]},{"t":"ul","items":[[{"text":"Somebody else was slightly better on the day, or knew someone on the panel."}],[{"text":"The internal candidate applied in week three."}],[{"text":"The headcount was frozen between your second and third interviews."}],[{"text":"The hiring manager left and the role was withdrawn."}],[{"text":"The team decided mid-process that they needed someone more senior."}]]},{"t":"p","runs":[{"text":"Companies rarely explain which of these happened, so candidates supply the explanation themselves, and the one people reach for is always the same and always the least likely."}]},{"t":"h2","id":"asking-for-feedback-so-that-you-sometimes-get-it","runs":[{"text":"Asking for feedback so that you sometimes get it"}]},{"t":"p","runs":[{"text":"Make it cheap to answer. A long email asking for a detailed review gets nothing; two lines gets a reply maybe a third of the time."}]},{"t":"quote","runs":[{"text":"“Thanks for letting me know. If there was one specific gap that decided it, I would find that genuinely useful: a single line is plenty. Either way I appreciated the process.”"}]},{"t":"p","runs":[{"text":"Two things to know. First, many companies have a policy against detailed feedback, on legal advice, because a specific reason can become evidence in a discrimination claim. Silence is often policy rather than disdain. Second, when you do get feedback, treat it as one data point. A single interviewer’s view is noisy; the same comment from three companies is a finding."}]},{"t":"h2","id":"the-log","runs":[{"text":"The log"}]},{"t":"p","runs":[{"text":"The tracker from the funnel lesson gets four extra columns for anything that reached a first conversation:"}]},{"t":"ul","items":[[{"text":"Stage reached.","b":true}],[{"text":"Their stated reason","b":true},{"text":", verbatim, or “none given”."}],[{"text":"Your own read","b":true},{"text":", written the same day, in one sentence."}],[{"text":"One thing to change","b":true},{"text":": and it must be one thing, not a list."}]]},{"t":"p","runs":[{"text":"Then review after every ten rows, not after every rejection. Reading it after each one is how a bad week becomes a theory about yourself. Reading ten together shows a pattern or shows that there is not one, and both are useful."}]},{"t":"p","runs":[{"text":"Patterns you will actually see: the same weak answer to the same question; falling out at the same stage every time; only applying when you feel confident, which correlates with roles below your level; four rejections all from companies whose process you disliked, which is a different problem."}]},{"t":"h2","id":"ghosting-and-a-rule-to-end-it","runs":[{"text":"Ghosting, and a rule to end it"}]},{"t":"p","runs":[{"text":"You will be ignored, sometimes after a final interview. It is rude, it is common, and it is structural rather than personal: a recruiter with 1,400 applications and a frozen requisition sends nothing."}]},{"t":"p","runs":[{"text":"Have a rule so it stops occupying you: "},{"text":"two follow-ups, one week apart, then close the row and stop thinking about it.","b":true},{"text":" Write the rule down before you need it, because the decision is harder in the moment."}]},{"t":"h2","id":"re-applying","runs":[{"text":"Re-applying"}]},{"t":"p","runs":[{"text":"Applying again after six to twelve months is completely normal. Their tracking system keeps your history, which cuts both ways: a strong previous process is remembered favourably, and six scattergun applications are also visible."}]},{"t":"p","runs":[{"text":"When a rejection says “please do apply for future roles”, it is sometimes politeness and sometimes literal. You can tell the difference by whether they name a role type or a timeframe."}]},{"t":"h2","id":"the-part-that-is-not-tactics","runs":[{"text":"The part that is not tactics"}]},{"t":"p","runs":[{"text":"Long searches are damaging, and the damage is not proportional to how tough anybody is. Three things reliably help, and none of them is more applications."}]},{"t":"p","runs":[{"text":"A fixed weekly quota","b":true},{"text":" (the eleven items from the volume lesson) so effort becomes a decision made on Sunday rather than a mood negotiated daily. On the days you do the quota, the day was a success regardless of any reply."}]},{"t":"p","runs":[{"text":"Other people.","b":true},{"text":" A study partner, a group, someone who is also searching. Isolation makes every rejection larger, and a search run entirely alone is the version that breaks people."}]},{"t":"p","runs":[{"text":"A separation you state out loud.","b":true},{"text":" The market’s response is information about the market. It is not a measurement of your worth, and the two feel identical from the inside after the fifteenth rejection. If it is affecting your sleep, eating or health, that is a signal to talk to somebody (a doctor, a friend, a helpline) and not a signal to send more applications."}]}],"check":{"question":[{"text":"A candidate has been rejected after three final-round interviews at three companies, with no feedback given. What is the most defensible interpretation?"}],"options":[{"text":[{"text":"Something specific about their interviewing is failing at the last stage and they should overhaul their approach before applying again."}],"correct":false,"tempting":[{"text":"Draws a strong conclusion from three trials of a stage where roughly two thirds of good candidates are rejected anyway."}]},{"text":[{"text":"The lack of feedback shows all three companies found the same serious problem and are avoiding saying it."}],"correct":false,"tempting":[{"text":"Reads a common legal policy against detailed feedback as concealment of a specific fault."}]},{"text":[{"text":"Reaching three final rounds indicates the funnel is working, three losses at roughly a one-in-three conversion is an ordinary outcome, and the useful action is to log each one and look for a pattern once there are ten data points."}],"correct":true,"tempting":null},{"text":[{"text":"They should apply to less competitive companies where they are more likely to convert."}],"correct":false,"tempting":[{"text":"Lowers ambition in response to what is statistically an expected run of results, and gives up the level they have demonstrably reached."}]}],"order":[1,2,3,0],"why":null},"prev":{"slug":"what-the-money-looks-like","title":"Pay, rates, and who actually decides them","minutes":8},"next":{"slug":"what-is-actually-in-an-offer","title":"The headline number is the least useful one","minutes":9}}}