No conduct flags in posts No reported consequences
Sentiment of their posts (what the content was like, not a character score)
Mostly positive+0.56 avg · 5 posts
negativeneutralpositive
3 positive 1 mixed 1 neutral
This profile is an AI reading of communication style and demonstrated behaviour in this person's own public posts
(prompt v4, rolled 2026-09-22 00:17:54). It is opinion about public content, not fact about the person,
and it never infers protected characteristics. Every score below links to the posts that produced it.
Style & behaviour attributes
Professionalism — register appropriate to a public/professional setting
90 · n=4
- “numbered systematic breakdown of six distinct ideas” — Sequoia Ascent 2026 summary
- “2025 has been a strong and eventful year of progress” — 2025 LLM Year in Review
Directness — says things plainly vs hedges or hides the point
87 · n=5
- “LLMs are no longer just chatbots or autocomplete” — Sequoia Ascent 2026 summary
- “the core issue is that benchmarks are almost by construction verifiable” — 2025 LLM Year in Review
Emotional regulation — measured and proportionate vs volatile or inflammatory
85 · n=2
- “I first started to internalize the shape of LLM intelligence” — 2025 LLM Year in Review
- “quite possible and desirable to train forward predictor” — Auto-grading decade-old Hacker News discussions with hindsig
Reliability signals — concrete commitments, follow-through, specificity over vagueness
85 · n=4
- “specific date: December 2025” — Sequoia Ascent 2026 summary
- “see DeepSeek R1 paper for examples” — 2025 LLM Year in Review
Consistency — claims and stance within the post are coherent
84 · n=4
- “verifiability thesis refined and extended” — Sequoia Ascent 2026 summary
- “Therefore, most of the capability progress of 2025 was defined by” — 2025 LLM Year in Review
Constructiveness — offers solutions, next steps or nuance vs only complaint
84 · n=4
- “My core automation framework is: traditional software automates what you can specify” — Sequoia Ascent 2026 summary
- “RLVR involves training against objective reward functions which allows longer optimization” — 2025 LLM Year in Review
Accountability — owns mistakes, gives credit, avoids blame-shifting
83 · n=2
- “I'm sorry I lost the reference to its original post” — 2025 LLM Year in Review
- “cost about $58 and somewhere around ~1 hour” — Auto-grading decade-old Hacker News discussions with hindsig
Respectfulness — how others, critics and opposing views are treated
79 · n=3
- “see DeepSeek R1 paper for examples” — 2025 LLM Year in Review
- “Give out awards for Most prescient and Most wrong” — Auto-grading decade-old Hacker News discussions with hindsig
Warmth — friendliness and care toward the audience or people discussed
70 · n=3
- “The following is a list of personally notable and mildly surprising” — 2025 LLM Year in Review
- “I spent a few hours browsing and found it very interesting” — Auto-grading decade-old Hacker News discussions with hindsig
Humour — uses humour; note whether it is at others' expense
60 · n=4
- “we are summoning ghosts” — 2025 LLM Year in Review
- “wait, Opus' code?” — Auto-grading decade-old Hacker News discussions with hindsig
Political stances 0 of 5 posts take one
No explicit political stance in the analyzed posts.
Tone
analyticalinstructivetechnicalwarmoptimisticthoughtfulhumorousdirectenthusiasticself-awareTopics
llmsartificial-intelligencesoftware-engineeringai-agentsllm-trainingreinforcement-learningai-capabilitiesbenchmarkingai/llm capabilitieshindsight analysisSources followed
Analyzed posts (5)
| Published | Post | Sentiment | Conf. | Own voice | Lean | Flags | Tone | Strongest attributes |
|---|---|---|---|---|---|---|---|---|
| Apr 30, 2026 | Sequoia Ascent 2026 summary
feed
Bulk of content is AI-generated summary/transcript (author transparent about this). Intellectual positions and ideas are author's own; prose is largely AI-assisted from his direction. |
positive +0.72 | 0.80 | 0.62 | — | — | analytical, instructive, optimistic | professionalism 90 · directness 85 · consistency 85 |
| Dec 19, 2025 | 2025 LLM Year in Review
feed
Post text ends abruptly mid-sentence; analysis based on complete content provided. |
mixed +0.65 | 0.85 | 0.95 | — | — | analytical, thoughtful, humorous | directness 85 · consistency 85 · emotional regulation 85 |
| Dec 18, 2025 | Chemical hygiene
feed
Description truncated mid-sentence; insufficient text to assess full argument structure, coachability, or emotional regulation. |
positive +0.35 | 0.40 | 0.95 | — | — | casual, self-deprecating, informative | directness 80 · warmth 70 · consistency 70 |
| Dec 10, 2025 | Auto-grading decade-old Hacker News discussions with hindsight feed | positive +0.85 | 0.88 | 0.95 | — | — | warm, enthusiastic, instructive | constructiveness 95 · reliability signals 95 · directness 90 |
| Nov 29, 2025 | The space of minds feed | neutral +0.10 | 0.85 | 0.98 | — | — | analytical, explanatory, comparative | professionalism 95 · directness 90 · constructiveness 85 |