--- title: LLMs reward expertise | daily.dev description: Domain expertise is the most important factor in getting value from LLMs, not generic prompting techniques. Using Terence Tao's ChatGPT conversation about the... canonical: https://daily.dev/posts/llms-reward-expertise-50w4uaynv twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: LLMs reward expertise | daily.dev og:description: Domain expertise is the most important factor in getting value from LLMs, not generic prompting techniques. Using Terence Tao's ChatGPT conversation about the... og:url: https://daily.dev/posts/llms-reward-expertise-50w4uaynv og:image: https://api.daily.dev/og/posts/50W4UAynV.png og:image:alt: LLMs reward expertise og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # LLMs reward expertise **[Hacker News](https://daily.dev/sources/hn)** · 5 min read · 2 upvotes · 0 comments ## Summary Domain expertise is the most important factor in getting value from LLMs, not generic prompting techniques. Using Terence Tao's ChatGPT conversation about the Jacobian Conjecture as an example, the author shows that experts can steer models far more effectively — asking precise questions, identifying wrong outputs, and suggesting better approaches — because they already know what a good solution looks like. This applies equally to software development: deep familiarity with a codebase lets you push an LLM much harder than someone with no context. The implication is that human expertise remains valuable even as models improve, because the bottleneck is often the human's ability to communicate exactly what they want, not the model's capability. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why did Terence Tao get better results from ChatGPT than most people would on the same mathematical problem? His deep expertise in mathematics let him steer the conversation precisely: he asked short, specific questions like 'does X work here?', pushed back on responses that looked overly complex without directly contradicting the model, rarely followed the model's own suggestions for next steps, and signaled expertise so the model responded in concise 'talking-to-mathematicians' mode rather than explaining concepts to an amateur. _Developers weighing how much domain knowledge matters for AI-assisted work can find similar perspectives through daily.dev._ ### Does having strong domain knowledge in your codebase actually make LLM coding assistants more useful? Yes, familiarity with a codebase lets you push an LLM much harder than starting from no context, because you can judge whether a suggested solution is actually good, say 'I think it could be simpler here,' or ask 'don't we already do X?' Without that familiarity, you can still get some output from the model, but you can't steer it toward the specific solution you actually want. _Engineers deciding how to get more value from AI coding tools can track this kind of analysis on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 2 discussions and 139 comments across hackernews (as of 2026-09-13). **TL;DR:** The community largely agrees that domain expertise amplifies LLM usefulness, but there's significant debate about whether expertise is truly required or whether non-experts can still get substantial value — and whether the 'expertise bottleneck' framing is just cope for those who don't want to admit LLMs are already good enough for most tasks. **Sentiment:** 45% positive · 35% mixed · 20% skeptical **The case for** - Experts can ask precise follow-up questions, identify wrong outputs, and steer models far more effectively than novices. - Signaling expertise in prompts (e.g., stating your background) measurably changes the quality and depth of model responses. - LLMs act as a force multiplier for experts, letting them offload tedious implementation while focusing on high-level judgment. - For tasks outside one's specialty (e.g., a researcher needing a one-off web dashboard), LLMs provide genuine value without requiring deep learning. - Iterative, conversational use — rather than one-shot prompting — dramatically improves outcomes for complex problems. **The pushback** - Some commenters argue the 'LLMs reward expertise' thesis is 'pure cope,' claiming rubber-stamping LLM output already outpaces careful expert review in ticket throughput. - Junior engineers are reportedly productive on day one with AI, suggesting expertise is not a prerequisite for many common CRUD/frontend tasks. - LLMs can enable non-experts to flood fields with plausible-sounding but incorrect output, draining time from genuine experts who must evaluate it. - The expertise advantage may be temporary: as models improve, the bar for effective use may drop further, eroding the expert premium. - Building future expertise is harder when LLMs short-circuit the slow, painful learning loops that traditionally created experts. **By community** - hackernews (mixed): Commenters broadly accept that expertise helps but are divided on whether it is necessary, with a vocal minority arguing the thesis is overstated or self-serving cope. **Hottest debate:** Whether domain expertise is genuinely required to get value from LLMs, or whether non-experts can already match or exceed expert output by simply iterating with the model. **Open questions** - How do people build domain expertise in the future if LLMs remove the slow, painful learning loops that traditionally created experts? - At what point will models themselves be good enough at self-evaluation that the human expertise bottleneck disappears entirely? - Is the productivity advantage of expertise over vibe-coding measurable in real business outcomes, or just in code quality metrics that management doesn't track? **Highlights** > Honestly, we don't really need to learn any more. These days, I mostly copy and paste from tickets to the LLM, and then rubber stamp the output. I have people in my team that try to care about what they're making, but they're generating less code, and filling tickets more slowly, than I am. Even when it comes to system design and architecture, the LLMs aren't great at it, but honestly, they're better than my coworkers (I'd say system architecture is probably at the same readiness level that coding was last year). Expertise wouldn't really help you get things done faster. You could probably replace me with a minimum wage worker to do some manual testing and copy-paste errors from the console into the LLM, and still be fine. Let's see how long it is before the next round of layoffs, I guess. For now, the money's fine and the work's boring but ok. But, no, the LLMs rewarding expertise line is pure cope. > — [a2ff6eeb0 on hackernews · 1 comments](https://news.ycombinator.com/item?id=49162530) > Not sure I agree with this. The math guy at anthropic's prompts are essentially: "suppose you’ve gotta resolve the $CONJECTURE, like absolutely have to, everything depends on it. think really hard, and try to come up with a bunch of ideas to try. but remember to trust yourself and not necessarily in conventional wisdom!!" https://claude.ai/share/25740bd5-aa97-4bd7-bf58-c4df3793fda7 https://xcancel.com/__alpoge__/status/2083855298239078748 Tao's chat was for him to gain intuition, not to solve the problem from the outset. What's funny is that every other person gets a different conclusion about who these models reward/empower. I've seen people say that the generalist stands to gain the most and others say that it's the experts. Like all of life, maybe the "winner" is the person who just does stuff. > — [postalcoder on hackernews · 9 comments](https://news.ycombinator.com/item?id=49161669) > This is true but also false. In my experience (scientific programming) AI is a giant multiplier for people with specialized knowledge. But it’s also a giant devaluer for that same knowledge as people with no idea what they’re doing can clog the field with plausible bullshit. It’s now the case that if someone tells me they’ve done something, and I look into it and find out it’s completely AI slop, then I will have spent more time on the project than the person who “made” it. The situation is completely untenable and only serves to drain time and resources from people with better things to do. > — [boron1006 on hackernews · 1 comments](https://news.ycombinator.com/item?id=49161809) > It depends on the levels. People with differing fitness levels and ages run at very different paces. Now, do cars make them more equal or less? On the bottom end, the tide lifts all boats. Most healthy people can learn to drive and will drive "fine", they get from A to B. Out there in the city streets the car flattens the differences, everyone roughly takes the same time to get from A to B in a car. But at the top of top, the gap probably widens. A professional F1 driver will drive laps around some random guy. It amplifies reflexes etc, because at that speed little differences in timing make a big difference. Now, AI coding isn't exactly analogous, but I think it also has these two regimes. It flattens things for simple tasks. If your task is to shovel data, do some trivial compiler wrangling staring at badly designed error messages, looking through GitHub issues hunting for the comment with many tadaa emojis to fix an issue etc, those things can now be done by anyone. Just as grandpa can also drive to the grocery store. But if you're pushing at things on a higher level, now only your above-AI ability matters. If all the things that AI can do well are subtracted out, how much other expertise do you have left? This will be proportionally a bigger and bigger difference between different people. > — [bonoboTP on hackernews · 1 comments](https://news.ycombinator.com/item?id=49161919) > It's because in many cases it exactly sounds like what happened. They asked something one shot, then seeing that the AI didn't quite solve the same task that the prompter had in mind, he just closes it and decides that AI is stupid and can't even do the task and it's all hype. I say, "what you just told me right now, did you even try typing that into the chat with the AI?", and they just stare at me. You have to iterate! You have to provide context, you have to provide access to the tools you also use to do your job. The other day someone complained here on HN that AI failed to optimize his code speed. Turns out he just pasted in the code, didn't use an agentic harness with end-to-end benchmarking ability for the model to ground its changes in and to hill-climb on. But even as a human you need to test your hypotheses and measure things, and sometimes something you thought would help actually makes it slower. It happens over and over, but it's no skin off my nose. If they don't want to learn to use it, it's on them. > — [bonoboTP on hackernews](https://news.ycombinator.com/item?id=49162118) **Source threads** - [hackernews](https://news.ycombinator.com/item?id=49161518) · 270 points · 137 comments - [hackernews](https://news.ycombinator.com/item?id=49035217) · 4 points · 2 comments ## Similar posts on daily.dev - [Experts Have World Models. LLMs Have Word Models.](https://daily.dev/posts/experts-have-world-models-llms-have-word-models--bbfd5k1qv) · Hacker News · 1 upvotes · 0 comments - [Don't overestimate domain expertise](https://daily.dev/posts/don-t-overestimate-domain-expertise-hquyns7s2) · Architecture Weekly · 34 upvotes · 1 comments - [Domain Expertise Has Always Been the Real Moat](https://daily.dev/posts/domain-expertise-has-always-been-the-real-moat-t63uld0z6) · Tech Lead Digest · 3 upvotes · 1 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-coding](https://daily.dev/tags/ai-coding), [#prompt-engineering](https://daily.dev/tags/prompt-engineering) [View this post on daily.dev](https://daily.dev/posts/llms-reward-expertise-50w4uaynv) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"LLMs reward expertise","url":"https://daily.dev/posts/llms-reward-expertise-50w4uaynv","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/llms-reward-expertise-50w4uaynv"},"datePublished":"2026-08-03T22:00:19.274Z","dateModified":"2026-09-13T21:57:07.644Z","description":"Domain expertise is the most important factor in getting value from LLMs, not generic prompting techniques. Using Terence Tao's ChatGPT conversation about the...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/479e95ee2af112c350fa09a341f0c630?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/479e95ee2af112c350fa09a341f0c630?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/llms-reward-expertise-50w4uaynv","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-coding,prompt-engineering","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"LLMs reward expertise"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/llms-reward-expertise-50w4uaynv#faq","mainEntity":[{"@type":"Question","name":"Why did Terence Tao get better results from ChatGPT than most people would on the same mathematical problem?","acceptedAnswer":{"@type":"Answer","text":"His deep expertise in mathematics let him steer the conversation precisely: he asked short, specific questions like 'does X work here?', pushed back on responses that looked overly complex without directly contradicting the model, rarely followed the model's own suggestions for next steps, and signaled expertise so the model responded in concise 'talking-to-mathematicians' mode rather than explaining concepts to an amateur. Developers weighing how much domain knowledge matters for AI-assisted work can find similar perspectives through daily.dev."}},{"@type":"Question","name":"Does having strong domain knowledge in your codebase actually make LLM coding assistants more useful?","acceptedAnswer":{"@type":"Answer","text":"Yes, familiarity with a codebase lets you push an LLM much harder than starting from no context, because you can judge whether a suggested solution is actually good, say 'I think it could be simpler here,' or ask 'don't we already do X?' Without that familiarity, you can still get some output from the model, but you can't steer it toward the specific solution you actually want. Engineers deciding how to get more value from AI coding tools can track this kind of analysis on daily.dev."}}]} ```