# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/yes-and-50003796 # Yes, and 666 points · 267 comments [Full discussion]() [Read original]() Category: [Industry & Society]() ## Skept-o-meter & Hotness Skept\-o\-meter: Low\. Estimated from 10 comments\. 17 comments for the summary\. Peak rank: \#1 Time in Top 10: 24\.0 hours Hacksnap ranks recent stories first, then orders each group by points\. Peak rank uses all retained observations\. Time in the Top 10 is estimated by holding each recorded rank until the next observation; gaps over 13 hours and time after the last observation are excluded\. Movement between observations is unknown\. 23 recorded rank observations from 2026\-10\-09T17:03:06\.673648\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\. Hotness — latest 23 recorded Hacksnap ranks: 2026\-10\-09T17:03:06\.673648\+00:00: rank \#3 2026\-10\-09T18:01:02\.851052\+00:00: rank \#3 2026\-10\-09T19:01:11\.250291\+00:00: rank \#2 2026\-10\-09T20:01:32\.91907\+00:00: rank \#2 2026\-10\-09T21:02:08\.376144\+00:00: rank \#1 2026\-10\-09T22:01:19\.253075\+00:00: rank \#1 2026\-10\-09T23:01:31\.990912\+00:00: rank \#1 2026\-10\-10T08:01:00\.593376\+00:00: rank \#1 2026\-10\-10T09:01:45\.527975\+00:00: rank \#1 2026\-10\-10T10:00:53\.799353\+00:00: rank \#1 2026\-10\-10T11:00:35\.65851\+00:00: rank \#1 2026\-10\-10T12:01:11\.169647\+00:00: rank \#1 2026\-10\-10T13:00:54\.008377\+00:00: rank \#1 2026\-10\-10T14:01:43\.208766\+00:00: rank \#1 2026\-10\-10T15:01:24\.825188\+00:00: rank \#1 2026\-10\-10T16:01:10\.213328\+00:00: rank \#1 2026\-10\-10T17:00:48\.84189\+00:00: rank \#38 2026\-10\-10T18:00:57\.55132\+00:00: rank \#38 2026\-10\-10T19:00:52\.379463\+00:00: rank \#37 2026\-10\-10T20:00:21\.993912\+00:00: rank \#37 2026\-10\-10T21:01:10\.657738\+00:00: rank \#37 2026\-10\-10T22:00:50\.433569\+00:00: rank \#36 2026\-10\-10T23:01:00\.95027\+00:00: rank \#37 Carson Gross argues AI makes programming fundamentals more, not less, important, but the discussion is split: reported productivity gains and comprehension debt raise doubts about whether juniors can safely skip writing\. ## The brief Carson Gross, a computer science lecturer, argues that programming remains a viable career despite AI because its core work—solving problems and controlling complexity—persists\. He warns juniors against using AI as a code generator, saying they must write code to learn to read it and avoid systems they cannot understand\. He rejects the assembly\-to\-high\-level\-language analogy because compilers allow predictable reasoning while LLM prompts require experimentation and can add accidental complexity\. AI is better used as a teaching assistant, he says, while future programmers should invest in communication, business understanding and architecture\. - Gross frames programming as problem\-solving and complexity control, skills he expects to remain valuable even as AI tools improve\. - He warns juniors against letting AI generate assignments, arguing writing code is necessary to read code and avoid systems they cannot understand\. - He rejects the assembly\-to\-high\-level\-language analogy: compilers allow predictable reasoning about source\-to\-behavior changes, while LLM prompt changes require experimentation and can add accidental complexity\. - He recommends AI as a teaching assistant for getting unstuck and has published an AGENTS\.md file to configure coding agents as tutors rather than code generators\. - For seniors, he endorses LLMs for code analysis, project organization, small snippets, regex and CSS, throwaway demos and test suggestions, but not full supported solutions or API design\. - He expects communication, business understanding and software architecture to gain relative value; the job market is weak but cyclical, with personal networks more effective than job sites\. ## Discussion themes Analyzed: 2026\-10\-09T19:00:56\.786876\+00:00 Analysis sample: Based on 19 of 35 usable stored comments. Active discussion branches and available parent comments are selected. The analysis input was further shortened to fit its context limit. This sample may omit parts of the full thread. Selected themes do not measure community opinion or how common a view is. ### LLM\-assisted coding as a middle ground for experienced developers Comments describe using LLMs for code generation or typing while retaining human reasoning, review, and system ownership\. One reply argues senior engineers are more effective because they keep a mental model and validate output; another commenter still writes code by hand and uses LLMs mainly for research and documentation\. Sources: [Comment 50013408]() · [Comment 50015214]() · [Comment 50016084]() ### Nondeterminism and maintenance risk in LLM\-generated code One thread argues prompt changes cannot be reasoned about with compiler\-like formal precision, so the only way to know an LLM's output is to run it\. Replies qualify this: nondeterminism matters less if engineers review and own the output, and some question whether developers really reason about code\. Separately, commenters worry that faster feature generation accumulates code faster than developers understand, while another reply expects AI to maintain features and developers to validate output\. Sources: [Comment 50012822]() · [Comment 50013408]() · [Comment 50013562]() · [Comment 50016603]() · [Comment 50013212]() · [Comment 50012198]() · [Comment 50013465]() ### Shifting developer roles and career advice One commenter expects fewer developers as LLMs improve, with the bottleneck becoming new revenue\-generating ideas, and would steer kids toward entrepreneurship using coding\. Replies debate this: one says software understanding helps generate ideas, another says domain experts outside software generate new ideas and may go to AI, and another predicts developer work shifts to writing specs, managing agents, and validating AI output\. Sources: [Comment 50012198]() · [Comment 50012323]() · [Comment 50012839]() · [Comment 50013465]() ### Security and untrusted LLM\-generated code A commenter says verifying that systems do what they claim safely and efficiently still requires technical knowledge, especially as security is under fire\. A reply points to Microsoft MXC as an example of running untrusted code such as model output\. Sources: [Comment 50020549]() · [Comment 50020605]() ### Economic viability of niche custom software One commenter proposes not scaling and instead serving a few small businesses at $2000 per month\. A reply disputes that clients will pay that for something they can now build in\-house, possibly better suited to their needs\. Sources: [Comment 50022197]() · [Comment 50022645]() ### Industrialization metaphor and custom versus mass\-produced software A thread compares hand coding to hand\-woven rugs displaced by mass production\. Replies question whether the artisan market shrank or mass production served new buyers, and one commenter finds the metaphor unclear for software\. Another argues software is inverse: one\-size\-fits\-all scaling frustrates users, so cheaper customization may serve smaller groups, while another says software base jobs repeat but implementation planning and flexibility remain human work\. Sources: [Comment 50019033]() · [Comment 50019125]() · [Comment 50019578]() · [Comment 50020549]() · [Comment 50019613]() ## Sources & coverage AI-generated summary · 2026\-10\-09T17:02:06\.923542\+00:00 Based on 17 of 32 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. The model input was further shortened to fit its context limit. Article text may also be shortened. Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.