Why tech workers are quietly dreaming of goat farms, and what it means if they actually go.
What the article says
- The essay argues knowledge work never had real meaning, it just borrowed a sense of purpose from busyness and status.
- AI is now doing the actual thinking and writing, which strips away even that illusion of being needed.
- The piece calls this Workism, a kind of secular religion where the office replaced church.
- It warns that losing faith in that religion, with no backup plan, is what is driving people toward pottery, farms, and other escapes.
- The author says the messy, collaborative parts of work were the real value, and AI is erasing exactly those parts.
What HN is saying
- Commenters compare this to printers losing their trade, but others push back that this shift could hit far more people far faster.
- Several say remote work, not AI, is the real source of their misery and isolation.
- One thread blames toxic online culture itself for the constant despair, arguing miserable people simply post more.
- A recurring theme is that people who got into tech for money rather than love of the craft are the ones burning out hardest.
- Some readers are frustrated the article ignores that AI coding tools have made parts of their jobs less stressful, not more.
A developer pushes back hard on the tech industry's favorite new put down of programmers
What the article says
- The writer is fed up with people saying coding was never the hard part now that AI can write it.
- He argues that claim insults programmers, pointing to years of high salaries, burnout, and thick textbooks as proof coding was always genuinely difficult.
- He is just as skeptical of the flip side, that figuring out what to build is the real hard part, since product people rarely get paid or scrutinized like engineers do.
- His actual point is that both understanding the problem and writing good code are hard, and pretending otherwise is just a coping mechanism.
- He closes by urging developers to broaden their skills but never hand over their judgment and taste to AI.
What HN is saying
- Commenters mostly agree the argument has a strawman problem. Nobody serious ever meant typing code was the easy part, they meant it compared to requirements, architecture, and people problems.
- Several programmers who work in enterprise or consumer apps push back the other way, saying sloppy code quietly wrecks products over time even if nobody notices at first.
- A recurring distinction shows up between writing code and understanding code, with people arguing the second is what actually takes skill and is what AI still struggles with.
- One commenter says the whole framing is unmeasurable, since crediting any single role for a project's success is basically impossible.
- A sharper take says the cliché mostly comes from organizations that were never willing to fund genuinely hard technical work in the first place, so of course their coding looked easy.
Denmark just made every high schooler defend their homework out loud, because AI made written work unverifiable.
What the article says
- Denmark now requires upper secondary students to orally defend any written assignment done at home, effective immediately.
- The rule targets about nine thousand students in the two year higher preparatory program who submit major papers each year.
- Schools are also told to add screen monitoring, restrict internet access during exams, and shift more assignments to supervised classroom time.
- Officials and student and teacher groups welcomed the move but called it a first step, saying policy will keep evolving as AI tools improve.
- Ministers want the changes to protect independent thought and academic skill, not just catch cheating.
What HN is saying
- Commenters note Denmark already uses oral defenses at the Master's level, so this feels like a return to old practice rather than something new.
- Several people say oral exams have deep roots across Europe, describing similar setups in Germany and Hungary where students draw a topic and get grilled by a panel.
- The sharpest pushback is about scale. Oral exams take a professor's full attention per student, so they work for small classes but become brutal or unfair once you have hundreds of students.
- Others raise accessibility concerns, since students with anxiety, speech differences, or language barriers face a harder time than they would with written work.
- A few teachers describe experimenting with their own workarounds, like quizzing students on take home material or requiring an audit trail of their AI use.
Oracle bans AI generated code from OpenJDK while Larry Ellison brags AI writes Oracle's own software.
What the article says
- Oracle has told OpenJDK contributors they cannot submit code generated by AI, citing safety, security and copyright risks.
- Developers can still use AI tools privately to debug or review code, just not to write what they submit.
- This clashes with what Oracle says about itself. Ellison recently claimed AI models now write Oracle's code, and a co-CEO credited AI with letting the company run smaller engineering teams.
- Oracle is pouring huge money into data centers this year, and a credit agency just downgraded its rating over doubts the spending will pay off.
What HN is saying
- Most commenters read this as Oracle protecting itself legally rather than making a quality judgment, since unclear copyright on AI code is a real liability for a project this widely used.
- Several point out the obvious hypocrisy of a company that publicly hypes AI writing all its own code while banning it from the code it gives away.
- One clarifying comment notes the ban seems aimed at outside community submissions, not necessarily Oracle's own core developers, so it may be more about limiting reviewer workload than a philosophical stance.
- A side debate breaks out over whether AI assisted coding, like autocomplete tools such as Cursor, counts as banned. Oracle's own FAQ says yes, even editing ten lines of AI output still counts as AI generated.
Every scrap of memory chip production through 2027 is reportedly already sold, mostly to AI companies.
What the article says
- Samsung, SK Hynix and Micron have reportedly sold out all their memory production through 2027, most of it to AI companies.
- Much of that supply was locked in through long term contracts, so new capacity years from now won't bring prices down quickly.
- Storage is following the same path. A basic terabyte drive that was cheap back in January now costs over fifty percent more.
- Consoles and PCs are already getting pricier because of it, and the writer says all they can do now is hope the AI bubble pops so RAM gets affordable again.
What HN is saying
- Commenters explain that AI chips eat up roughly three times the factory capacity of ordinary memory chips, which is why they crowd out regular RAM.
- Plenty of people are just venting. Broken PCs, no good upgrade path, and parts costing double what they did a few years back.
- The sharpest disagreement is whether this counts as real inflation or just normal supply and demand, and whether ordinary buyers will even feel it since most people don't replace devices every year.
- Several commenters pin their hopes on China's CXMT ramping up its own memory production as a way around the shortage.
- One person notes this already happened with graphics cards during the crypto boom, and expects the industry to shrug this off the same way.