2026-08-25T06:33:33Z https://nostr.now Robert Roskam https://nostr.now/npub1ul9r350gnhtsgk2cwc6uemuh206es9xp9wzwa5dgyf3hperaxcuq9cf0d4 https://files.mastodon.social/accounts/avatars/000/213/939/original/1a8ae941d7b9ac95.png https://files.mastodon.social/accounts/avatars/000/213/939/original/1a8ae941d7b9ac95.png https://nostr.now/nevent1qqsdqkgqltxd7wgmkqu4dc4z6etsnd9lu9nn39h0fhjjxjnmtzw58yqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsyhpspn Hey folks, about a year ago I started to write a book. I found myself over the past year wanting to actually quote things I had written fairly often, and send links to people.<br/><br/>So while I continue to write the longer book, I decided I should publish a brief portion of what I&#39;ve written.<br/><br/>I hope you enjoy!<br/><br/><a href="https://robertroskam.com/books/maxims-of-software-engineering/">https://robertroskam.com/books/maxims-of-software-engineering/</a> 2026-07-12T13:41:57Z https://nostr.now/nevent1qqsrftkdz7h450l3lagyhk0s6h5gsv8z3pjrnnjz22my978tucjuhlczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrse68q7x Software principles that last don&#39;t come from architects in glass offices. They come from practitioners who noticed what kept working.<br/><br/>These principles arose from doing the work. They did not descend from on high from those far removed from the work.<br/><br/>This, I feel, is the most fundamental truth about doing work: those without dirty hands are wrong. Doing something makes you right. 2026-07-08T22:46:30Z https://nostr.now/nevent1qqs8tk4vcl965r7y5cxm6uvuk7jtugs8kvh78tadjtg6538qj8wtueszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrswjc7mv A pair of bangers.<br/><br/>The data black hole at the center of AI by Dwarkesh Patel<br/><a href="https://youtu.be/4pG3SJQPAwk">https://youtu.be/4pG3SJQPAwk</a><br/><br/>AI is turning workers into tools. The Atlantic with <span itemprop="mentions" itemscope itemtype="https://schema.org/Person"><a itemprop="url" href="/npub1fdrpydz4n2dpgpy5xhsa6rjg8ytggn82mt0addw6sf93272nhhnqsjlvhs" class="bg-lavender dark:prose:text-neutral-50 dark:text-neutral-50 dark:bg-garnet px-1"><span>Cory Doctorow</span> (<span class="italic">npub1fdr…lvhs</span>)</a></span> <br/><a href="https://youtu.be/SPQNPJ0CEPo">https://youtu.be/SPQNPJ0CEPo</a> 2026-06-19T22:19:37Z https://nostr.now/nevent1qqsdektx4z2997r2telgsku5z5hak0kwl4yv6k676v43xpa06wt8z8gzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsua590d 🔥 take for this audience. LLMs are a dead end for AGI, a dead end for many applications, and that still makes them useful.<br/><br/>They&#39;re just not the genies that certain financially incentivized people claim they are.<br/><br/>The big model makers are incredibly wasteful and incredibly not transparent. Why would they be? Their ability to make money hinges on this stuff seeming to be magic.<br/><br/>It&#39;s not magic. It&#39;s math. We&#39;ve taught old rocks new tricks.<br/><br/>1/ 2026-04-23T03:08:05Z https://nostr.now/nevent1qqsv8t5q9ws769xpvz03dnjnfdx3ctmfj4tzm5tjn0dvt45jd84gveqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs9j0zzs This is a documentary <br/><br/><a href="https://youtu.be/xE9W9Ghe4Jk">https://youtu.be/xE9W9Ghe4Jk</a> 2026-03-07T07:55:41Z https://nostr.now/nevent1qqsz95utv79qzhxefygqjcwcgurpq7y8gt8x8t062pqsj9vke8k747czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsmg72vw In reply to <a href='/nevent1qqsgtlgf9ksznu50lthkc3anefsqnam3r7nds40xd70qd040rlz20tcq2dspa'>nevent1q…dspa</a><br/>_________________________<br/><br/>It&#39;s one thing to have a deadline, to have pressure.<br/><br/>And it&#39;s another to have to do that while you&#39;re being yelled at to go faster or to have your favorite tools/processes be taken away and be given the cheapest possible ones that break constantly.<br/><br/>Those aren&#39;t constraints. They&#39;re a framing of work as mere tasks—a removal of joy, and even a hostility to it.<br/><br/>3/ 2026-02-28T18:49:30Z https://nostr.now/nevent1qqsgtlgf9ksznu50lthkc3anefsqnam3r7nds40xd70qd040rlz20tczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsc5sq4s In reply to <a href='/nevent1qqsgxqea38ugjjs86wzu8gn0sl9w0rlf5l6lwhp50nlayefhq4qeaycg2anx2'>nevent1q…anx2</a><br/>_________________________<br/><br/>Creativity often thrives in constraints. Time, space, material, method, or tools. Removing or restricting these sometimes leads to amazing results and experience. Less is definitely more very often.<br/><br/>Constraints are one thing, but hostility is another.<br/><br/>2/ 2026-02-28T18:49:22Z https://nostr.now/nevent1qqsgxqea38ugjjs86wzu8gn0sl9w0rlf5l6lwhp50nlayefhq4qeayczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsrw9ll6 There is an assumption that joy and work are different things.<br/><br/>If something is extremely boring or annoying with a task at work, it&#39;s treated as an &#34;oh well, that&#39;s life&#34; by most people and most bosses.<br/><br/>I think that&#39;s fundamentally flawed.<br/><br/>I also don&#39;t think joy means to make things easy. Some of the most delightful and creative experiences I&#39;ve ever had are when I had to work in constraints.<br/><br/>1/ 2026-02-28T18:49:04Z https://nostr.now/nevent1qqsr76nmfqhte73r8d7g835zmu9nmylhz9nmd28g6zszkkfs27ezrzszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsqjm9dj Software principles that last don&#39;t come from architects in glass offices. They come from practitioners who noticed what kept working.<br/><br/>These principles arose from doing the work. They did not descend from on high from those far removed from the work.<br/><br/>This, I feel is the most fundamental truth about doing work: those without dirty hands are wrong. Doing something makes you right. 2026-02-01T14:01:13Z https://nostr.now/nevent1qqs9djpegx9f9zt6he4q6r7hqxeqekwgv54vxd2tmla7dx298l2yvhgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs7f45wp You&#39;ve been told software engineers &#34;write code.&#34; But it&#39;s not true.<br/><br/>If I—a bog standard average software engineer—were to screen records myself &#34;writing code&#34;, and then play it back for you. <br/><br/>You&#39;d notice something. <br/><br/>I do barely any typing. I&#39;m switching between files, repositories, documentation, tabs, terminals, IDE tabs. What am I doing? <br/><br/>I&#39;m trying to bring it altogether in my head. That&#39;s where the action actually happens first before it makes it out of my fingers.<br/><br/>1/ 2026-01-31T14:01:09Z https://nostr.now/nevent1qqsf2jwyqetvp3kmfrdw09fgakh44cvdrzramqs0ugtl3f9rj30eurqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsy7q40h The stages of debugging a race condition:<br/> <br/><br/> 2. anger<br/> 4. depression<br/> 3. bargaining<br/> 5. acceptance<br/> 1. denial 2026-01-04T14:00:51Z https://nostr.now/nevent1qqs87hnj5e440v054l4dm6yarj8s7l3fa6y2wlcjpafagaxnm0rnxuqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs66jkz8 There are only two options, as an engineering manager:<br/> - you&#39;re the constraint<br/> - you&#39;re finding the constraint<br/> <br/>Any kind of leader, but especially in technical roles, is tempted to do the work.<br/> <br/>If you&#39;re routinely doing the work, there&#39;s no one else doing the things that only a manager can &amp; should do.<br/> <br/>The main thing you should be doing is ensuring that all of your direct reports aren&#39;t blocked. <br/> <br/>When they&#39;re waiting on you, then you are the constraint. 2026-01-02T14:01:32Z https://nostr.now/nevent1qqsx4jupsy4gr2ztw9crc3vw8panmz66v74kfcun0pk3qpulhp9tgjqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsw6z2tq YouTube demonetized Patrick Boyle&#39;s video on the Epstein Files. Which effectively hides it over there.<br/><br/>I&#39;m doing my part to ensure his hard work doesn&#39;t go to waste.<br/><br/>Watch his video and reshare.<br/><br/><a href="https://youtu.be/GAJf2F1BbRA?si=ZqE1jVr1oMFpGeOM">https://youtu.be/GAJf2F1BbRA?si=ZqE1jVr1oMFpGeOM</a><br/> <img src="https://files.mastodon.social/media_attachments/files/115/812/193/599/951/802/original/53b6db9117b774c0.png"> <br/> 2025-12-31T04:06:18Z https://nostr.now/nevent1qqs8uzkfjy9tw4r5stztcuykj48q2x6g0wefjgtztzvmddkl8pj3fzczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsccc9p6 TIL that Speilberg did not direct Back to the Future. #mandelaeffect much? 2025-12-24T15:48:15Z https://nostr.now/nevent1qqsvep56ym43c6c0ma64ple224quhzr0l3t5hsz9kxfd8qa3e4uasfgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsdrzrzl Functional programming won&#39;t solve all your problems. If you insist on it, you&#39;ll just move them around on your plate. <br/><br/>Was recently re-reading the Unicorn Project and the main character had some wildly optimistic views about functional programming that I&#39;ve heard echoed by others in real life.<br/><br/>Sometimes you want to model your problem space with shared state. If you&#39;re like, &#34;nah, I only do functional stuff&#34;, you&#39;ll find yourself doing a lot of extra work ime just to do the same thing.<br/><br/>1/ 2025-12-18T14:00:33Z https://nostr.now/nevent1qqs06cedpv9q85uxgmfyz0mwc2ygxan3l2k9xtxqwqsyxqg2ndyfuaszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsnspjrq Anti-goals for software engineering: <br/><br/>- using AI for the sake of it<br/>- feeling clever while writing it<br/>- writing it as short as possible; 1 liners vs 10 liners<br/>- covering every potential use case<br/>- adding abstractions that you might need<br/>- optimizing the performance of every function<br/><br/>That&#39;s definitely not a complete list. What would you add to it? 2025-12-09T14:00:33Z https://nostr.now/nevent1qqs05sg5c294kt27qs3k36dcd3shhxcz4z027jfc50rnfw80pke7slszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsvrgxyl Anti-patterns for distributed services:<br/>- No timeout handling<br/>- Missing retry logic<br/>- Synchronous calls in critical paths<br/>- Assuming 100% service availability<br/>- No circuit breakers<br/>- Complex distributed transactions<br/>- Ignoring backpressure<br/>- Treating remote calls like local ones<br/>- Hardcoding IP addresses<br/>- Assuming cloud regions never fail<br/><br/>What should I have added? 😱 2025-12-07T14:00:23Z https://nostr.now/nevent1qqsp6a6kcwytpv9q22fcskz29ycnk8tls69mzskxvrudg7dwtm53gmszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsx9tavp LeetCode is popular, not because it works. Nah, it&#39;s for another reason.<br/><br/>It makes the interviewer feel good.<br/><br/>There&#39;s comfort in having a &#34;objective&#34; measure. A binary pass/fail or even a grade. It reduces the anxiety of making a hiring decision, makes it seem closer to &#34;fair&#34;.<br/><br/>We choose the metric that&#39;s easiest to measure, not the one that&#39;s most predictive of success. Then we convince ourselves it&#39;s the right metric because it&#39;s the one we have data on. 2025-12-06T14:00:27Z https://nostr.now/nevent1qqspclrrr5clyhve7cynhj6knt958u8jr92gpz9da5dnaq8grxxxp9szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs35hcqm Users do not care about:<br/>- your pitch deck<br/>- your fancy JIRA process<br/>- how many 5-min calls you hop on<br/>- how many story points that last ticket was<br/>- your runway<br/>- if you use kubernetes, serverless, or bare metal<br/>- if you code in Python or Javascript or C# or Java or Cobol<br/>- if your code coverage is 1% or 100%<br/><br/>They care about software that solves their problems.<br/><br/>All those things are things should care about and have an opinion on. But they all come second to software that solves their problems 2025-11-24T14:01:13Z https://nostr.now/nevent1qqs93aq7csjk5j0v6h6rv2may83j2qzyyqcc8unrqwgn5rdehr73yfqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsk4qpzf Discovering you&#39;re wrong &amp; changing is an underrated superpower. 2025-11-16T17:14:11Z https://nostr.now/nevent1qqsg7ywkgf4ugt062jat2kkuhaqdxqwcmw7aju3t8rqu6sle69m5n7czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsv0gd7d The fables of 10x engineers never include:<br/>- updating the docs<br/>- pair programming<br/>- adding meaningful test coverage<br/>- listening to burned-out coworkers<br/>- improving on-call runbooks<br/><br/>The engineers that do these do not seek glory. They plant trees they don&#39;t expect to sit under. 2025-11-16T03:18:16Z https://nostr.now/nevent1qqsw73u9wqnsv0xde029gqldeq3m2205f42ug4c9fgxnmtf8snudqsszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsd75ge9 I wish more software engineering leaders remembered that typing out the code was never the hard part. <br/><br/>Typing out the right code. That was the hard part.<br/><br/>AI doesn&#39;t help with that. AI tools will happily:<br/>- generate 5k line functions<br/>- install 100 packages to handle something built into a language<br/>- optimize for non-bottlenecks<br/>- write 10x tests for the same 5 functions and leave most untested<br/><br/>AI is cool, but you still need brain on, so your AI buddy doesn&#39;t walk you into a dead end. 2025-11-07T16:01:13Z https://nostr.now/nevent1qqspyrk2v7w3mrxr2g4yjnwxa0vgcc48htceeztdrvukjnsnevkt4lszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs4vczvz &#34;It depends&#34; is the best universally correct answer for technical discussions.<br/><br/>It&#39;s closely followed by:<br/>- throw more hardware at it<br/>- use a monorepo<br/>- add caching<br/>- refactor something <br/>- use microservices <br/>- remove caching <br/>- rewrite it<br/><br/>What did I miss? 😅 2025-10-31T12:00:56Z https://nostr.now/nevent1qqsy73c9c557ea4sgslxswj8jawz8c2qsaqlkqefl62xrdw7mgufk8czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs9r8mep I wish more software engineers knew about my favorite debugging method, bisection.<br/><br/>I usually call it the &#34;cut it in half&#34; approach. As an aside, I seem to recall learning it when in college and using it as a method to identify shorts in an electrical circuit.<br/><br/>1/ 2025-10-20T12:00:58Z https://nostr.now/nevent1qqs0eh99sqq05wsvv86zpnztnupjkp46tfmy5ckaxt9na7f9c3jketgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsgnfyz2 I wish more software engineers understood Amdahl&#39;s Law.<br/><br/>&#34;The overall performance improvement gained by parallelizing a computation is limited by the sequential portion of the computation.&#34;<br/><br/>Or stated a little less formally, only speeding up at the bottleneck of the system matters.<br/><br/>1/ 2025-10-14T12:00:57Z https://nostr.now/nevent1qqsvjxfrpw79kn5sk9s26c4qg6ydza5d63vxs88zk8q4xmvltv0k55czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsdu74z7 How many cups (8 oz/240mL) of caffeinated drinks/coffee/tea do you drink during a typical workday? 2025-09-28T22:00:42Z https://nostr.now/nevent1qqstsmp4grpx4975lh966j3vlmua5pl4x82h8e5mynj6cuc8edmh0sqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrstxkxkf What&#39;s your most productive time of day? 2025-09-21T22:00:23Z https://nostr.now/nevent1qqsrlla3nx9rp20tm45jv2e9j80u75ahsu5987g0drv9zcja4lzpf6czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrscjt5fh Good principles for making software seem to:<br/><br/>- Come from making software<br/>- Urge simplicity<br/>- Suggest pragmatism<br/>- Tend to brevity<br/><br/>Anything that doesn&#39;t include these tends towards specification rather than philosophy. 2025-09-14T12:00:12Z https://nostr.now/nevent1qqsg5ppvu8p2agrty6ndxyu4lch6kt0skp2rltmyekxmlz4h6xuu70czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrssgugzw I&#39;m not anti-LLM. I&#39;m just pro-thinking.<br/><br/>Everyone I know who is REALLY into AI uses it for reading EVERYTHING longer than 1000 words.<br/><br/>It&#39;s the mental equivalent of eating only smoothies, because they go down easier.<br/><br/>I say this with love: reading is thinking.<br/><br/>Stop letting slop think for you. 2025-09-09T12:00:47Z https://nostr.now/nevent1qqs0yc6jgj6mpx5pcusfjhlw0nqge00adwehea663s0x2yzh0mumg4czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsgcvmlz I&#39;m not anti-AI. I&#39;m just pro-thinking.<br/><br/>What do I mean? I mean that everyone I know who is REALLY into AI uses it for reading EVERYTHING longer than 1000 words.<br/><br/>This is the mental equivalent of eating only smoothies, because they go down easier.<br/><br/>I say this with love: reading is thinking.<br/><br/>Stop letting slop think for you. 2025-09-03T12:01:08Z https://nostr.now/nevent1qqsvqwz8qzvggjh2rff2gzml3em8t4rm2j94uwurycas94wzfxlv5tgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrss9s4ra Good news everyone!<br/><br/>You still don’t have to use JIRA in order to do #agile. 2025-09-02T12:01:07Z https://nostr.now/nevent1qqs2qdj8t5u4xzf0pxpvchw7takhx3kccgstvra6rulhthpazjqd9yczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsc5mhp2 |￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣￣|<br/>| Get off social media and hug someone |<br/>|＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿|<br/> \ (•◡•) / <br/> \ / <br/> ——<br/> | |<br/> |_ |_ 2025-08-30T12:01:02Z https://nostr.now/nevent1qqszuvkmkel006dc63z47d9wgczu0lduq8ltp5a8efskvfakg0p7a9szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsmlh373 Birthday Problem: In a group of 23 people, there&#39;s a 50% chance that two people share the same birthday.<br/><br/>This counterintuitive probability applies to software engineering:<br/>- Hash collisions happen more often than expected<br/>- Database ID collisions occur sooner than you think<br/>- Cache key conflicts emerge with fewer entries than anticipated<br/>- Bug report duplicates appear with smaller user bases<br/><br/>The lesson: when dealing with random events, our intuition about probability is usually wrong. 2025-08-21T12:04:30Z https://nostr.now/nevent1qqsxqmn9m3udyxpf0a06rmlg5hkq3h0pldq6ndgl67k0eg2y4hq4cygzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrstmmlac Principle of Least Astonishment: A system should behave in a way that least astonishes its users.<br/><br/>In software engineering:<br/>- APIs should work the way users expect them to work<br/>- Error messages should be clear and actionable<br/>- Configuration should follow common conventions<br/>- User interfaces should behave consistently<br/><br/>When your system surprises users, it&#39;s usually a design problem, not a user problem. Optimize for user expectations, not implementation convenience. 2025-08-18T12:01:00Z https://nostr.now/nevent1qqsg22cj466s2v8r9s3yawnk7sd27dyzgcy4pkaprcxs42jrjkr6a9qzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsw3r6gr Productivity in software dev has one real measure: is useful stuff getting to the users? <br/><br/>Everything else is misleading and/or gameable. 2025-08-17T13:00:59Z https://nostr.now/nevent1qqsxpqrgfl948tg4pae2f9n2lwpyakfwwzvqwvgf0l89m2tgxa3vsdgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrstugp77 Assumptions software engineers make about production networks that are often wrong:<br/><br/>- Network requests always succeed<br/>- Networks are reliable<br/>- Bandwidth is infinite<br/>- Latency is zero<br/>- The network is secure<br/>- Topology doesn&#39;t change<br/>- Nothing is shared<br/>- The administrator is trustworthy<br/>- Transport cost is zero<br/>- The network is homogeneous<br/>- Clocks are synchronized<br/><br/>These assumptions, known as the &#34;Fallacies of Distributed Computing,&#34; lead to fragile systems that break in production. 2025-08-13T12:00:27Z https://nostr.now/nevent1qqswh00vzcppvpqcpe895yau30g3346jftn0tqnh7ng4hptuwj93mnczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs0agn94 Lies programmers believe about scaling: <br/> <br/>- if 1 process can handle X requests, then 2 processes can handle 2x requests <br/>- well, 2x requests with 2 processes is possible if there&#39;s more than 1 CPU core <br/>- 2x reqs with 2 procs is possible with more than 1 CPU core and plenty of RAM <br/>- uh...and there&#39;s plenty of disk space? <br/>- oh, oh, right, and if there&#39;s no shared services <br/>- something, something Kubernetes <br/>- ...this can&#39;t be this hard <br/>- I&#39;ll have time to figure it out later 2025-08-12T03:36:37Z https://nostr.now/nevent1qqsfzl2gtrvwze94q345r48nssyw639r8w6eqj4nz85tedrgsma0u9qzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsqvls34 Moved my blog from Substack to self managed:<br/><br/><a href="https://robertroskam.com/blog/">https://robertroskam.com/blog/</a><br/><br/>I flipped the switch in the past hour or so.<br/><br/>I expect it’s still going to be bare bones for a bit. <br/><br/>Enjoy! 2025-08-10T22:08:54Z https://nostr.now/nevent1qqsz94r4ftml5zn2deydrnsn6n0llwz0ulzp6zucl77dv225eqev66czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrse65yng Things no one will mention at your funeral:<br/><br/>- participated in standups<br/>- increased to shareholder value<br/>- got in expense report on time all the time<br/>- had 90% code coverage on all PRs<br/>- clocked your time on time all the time<br/><br/>This is not an exhaustive list by any means. 2025-08-10T12:00:19Z https://nostr.now/nevent1qqst4zmnv8kkfd32lelq8g7znwmtec2vy7tsjgvr69nj3tdll8gw3sgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs0wyq7m The 3 universal constants in programming:<br/><br/>1. The speed of light<br/>2. It is more complicated than you think<br/>3. &#34;It works on my machine&#34;<br/><br/>Scientists are still determining which one is more fundamental 2025-08-08T12:00:26Z https://nostr.now/nevent1qqsp2wmr2pq2743rtx5zf3zf499rj3nqv4mvhm5k73qhh8cz93gwntqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrstyt3jp If you want to be a 10x engineer:<br/>- Share your ideas freely with little concern for credit<br/>- Default to openness<br/>- Hear another team member all the way out, especially when you disagree<br/>- Fail together, not alone—no one learns if you keep it secret<br/>- Message where you&#39;re going as much as current location<br/><br/>The 10x engineer lifts up their team to get 10x things done. 2025-08-06T12:00:48Z https://nostr.now/nevent1qqsfm8w4e7n0t7cm9enqlyluypwewg69dqe53x7jl88mam9l5v8zvpqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsj46urn Internet programming attacks through the ages<br/><br/>80s: hehe...I pretended to be Bob on IRC<br/>90s: 1k people went to my website at once and the server caught on fire<br/>00s: PHP/ActiveX/Flash - we make tech that&#39;s basically designed to be hacked<br/>10s: one of these 10k nodejs dependencies got hijacked, glhf figuring out which one<br/>20s: the LLM got prompted-injected via emojis 2025-08-05T12:00:40Z https://nostr.now/nevent1qqsx50uhxvv2vnd7y2qqanhhnvdynqn2mz0vltrx7jrpv4k624mwdzczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsg390vc Hofstadter&#39;s Law: It always takes longer than you expect, even when you take into account Hofstadter&#39;s Law.<br/><br/>This law captures the fundamental difficulty of software estimation. Even when we know we&#39;re bad at estimating, we&#39;re still bad at estimating.<br/><br/>- The discovered work emerges because...well, you &#34;discovered&#34; it<br/>- Scope creep happens gradually<br/>- Integration takes longer than expected<br/><br/>The solution isn&#39;t better estimation; it&#39;s shorter feedback loops. 2025-08-03T12:00:17Z https://nostr.now/nevent1qqsxdgcvd9vp4agrateaeeas3fzq2hq9e2xa8x7mnnyf0lzn65whleqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrssm2pn9 Make each tool do one thing well is still good advice.<br/><br/>We don&#39;t need any software to be a Swiss Army Knife, a bunch of bad tools stuffed together into the same tool. We do not have a space constraint with software tools.<br/><br/>In software, if you put lots of tools together, you make it confusing and hard to use with other tools. <br/><br/>You don&#39;t wind up with multiple good tools. You have a bad one. 2025-08-01T12:00:25Z https://nostr.now/nevent1qqs0zreq4wtf6z5s8ljszjnm2c25u8d3gp8n5lyxyjxyf5uar6d8reszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs6pugqu The best software engineers solve problems, not just write code.<br/><br/>They ask:<br/>- What problem are we actually trying to solve?<br/>- Is this the right problem to solve?<br/>- What&#39;s the simplest solution that could work?<br/>- What are the trade-offs?<br/>- How will we know if it&#39;s working?<br/><br/>Writing code is the implementation detail. Understanding the problem space, considering alternatives, and thinking through implications—that&#39;s where the real value lies.<br/><br/>Code is just the tool. Problem-solving is the skill. 2025-07-31T12:00:25Z https://nostr.now/nevent1qqstz9h4e8j5j5tknfl53c34llzej9j6f78qecvggpnj8kltlrut4fszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrscdv3wd Generally speaking, learning things entirely by yourself is the hardest way.<br/><br/>Knowledge alone isn&#39;t enough.<br/><br/>You need guidance to bridge the gap between theory and practice. <br/><br/>This is true whether you&#39;re learning:<br/>- to ride a bike<br/>- to code<br/>- to play an instrument<br/>- or any other valuable skill<br/><br/>The path to mastery is rarely found in books alone. It&#39;s found in practice with feedback.<br/><br/>Wisdom lives in the gap between knowledge and skill. 2025-07-30T12:00:16Z https://nostr.now/nevent1qqsycdx5trqu89ryy4n6a8j5h8ngnu3wja58camd9emnq25ygpjay0szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsx7tuzg In reply to <a href='/nevent1qqsfad7xrkhvq25a203c9dhuxcg5n06p9znr4k5x0kw8umyggheuvrg3t6pgg'>nevent1q…6pgg</a><br/>_________________________<br/><br/>yep! The original thing is called Parkinson&#39;s law of triviality. 2025-07-23T02:10:14Z https://nostr.now/nevent1qqsygxs27zue4pmhtrwwueeg3cysnh2fvrgh4wg5ncnez5krckjyt2szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs5cgqq8 Bike Shedding: &#34;People give disproportionate attention to trivial issues while ignoring important ones.&#34;<br/><br/>This appears everywhere in engineering:<br/>- Code reviews that focus on formatting while missing logic bugs<br/>- Architecture meetings that debate framework choices but skip scalability planning<br/>- Sprint planning that argues about story points but ignores technical debt<br/><br/>Complex topics get less discussion because they&#39;re harder to understand. 2025-07-23T01:00:14Z https://nostr.now/nevent1qqs2vjws8t0pmqdmfln8arstj2eua6nr3wdegk0hrw9gvarh34w94fqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs8fv395 Software engineering is about creating systems to deliver value. <br/><br/>Some work is direct (&#34;change button text&#34;) while other work is indirect (&#34;decrease error rates&#34;). <br/><br/>Many engineers would say this indirect work—building abstractions to enable rapid delivery—is the real work.<br/><br/>I&#39;m on the fence. What do you think? 2025-07-22T13:01:04Z https://nostr.now/nevent1qqszqdac7lgxvrqjzaav95uptnapt883spwja6rqk0hpklwc23lcmjczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsw57ayd Kernighan&#39;s Law: Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it.<br/><br/>Write code for the person who has to debug it at 2 AM. It&#39;ll likely be you. 2025-07-17T13:01:06Z https://nostr.now/nevent1qqsx58xx350ls7ysqjpj7qtgdtxuj85wmfqukjn0z8ca3uv8kgyd40qzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs0f2su2 Conway&#39;s Law: Organizations which design systems are constrained to produce designs which are copies of the communication structures of these organizations.<br/><br/>You can see this everywhere:<br/>- Monoliths built by small teams<br/>- Microservices proliferating in large organizations<br/>- Awkward interfaces that mirror organizational boundaries<br/>- Redundant systems when departments don&#39;t communicate<br/><br/>Software is made by groups of humans, and social structures provide defaults to technical decisions. 2025-07-12T13:00:26Z https://nostr.now/nevent1qqsyfyadlw89nptg0qpefv86rgfhpguva6qw3aurnmhvq94e3e09sdszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs7043px I died laughing on the first question.<br/><br/><a href="https://jsdate.wtf/">https://jsdate.wtf/</a> 2025-07-12T04:15:55Z https://nostr.now/nevent1qqs87ap3343fgdg22g99xzrl87fnp7pxynpx367n8xulj3jdpe56ecgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs9xayah Birthday Problem: In a group of 23 people, there&#39;s a 50% chance that two people share the same birthday.<br/><br/>This counterintuitive probability applies to software engineering:<br/>- Hash collisions happen more often than expected<br/>- Database ID collisions occur sooner than you think<br/>- Cache key conflicts emerge with fewer entries than anticipated<br/>- Bug report duplicates appear with smaller user bases<br/><br/>The lesson: when dealing with random events, our intuition about probability is usually wrong. 2025-07-12T01:00:03Z https://nostr.now/nevent1qqsx9lum9yyd9m09xt8rg693k9hgt5psp3uswd882p3820q8mshxgagzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsz36sc9 &#34;Technical Debt&#34; is a terrible description.<br/><br/>The metaphor suggests debt that can be &#34;paid off&#34; through dedicated effort. But most of what we call technical debt isn&#39;t debt at all.<br/><br/>Better metaphors might be:<br/>- Technical erosion - gradual degradation requiring maintenance<br/>- System friction - resistance that slows development<br/>- Legacy investment - working systems that need careful modernization 2025-07-11T13:00:56Z https://nostr.now/nevent1qqspqzehku3y2k9dj837nae554ejj2upe7nlgn832mrzxpc8w4mca2qzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs6m8ysh Knuth&#39;s Law: Premature optimization is the root of all evil.<br/><br/>But the full quote is more nuanced: &#34;We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%.&#34;<br/><br/>The wisdom isn&#39;t &#34;never optimize&#34; but &#34;optimize at the right time with the right data.&#34;<br/><br/>Optimize when you have performance problems, data showing bottlenecks, and understand the cost/benefit trade-offs. 2025-07-11T01:00:33Z https://nostr.now/nevent1qqsxwr8j2jt8q4ycs8muv3lnvwc7ejm4786n9eechsrrd6rsjjrx3mqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsxzd69v AI is hurting engineers.<br/><br/>Many new engineers think the AI is smarter than they are. It&#39;s not.<br/><br/>You&#39;re paid to think about the problem. One important way you learn to think about the problem is to go through the exercise of writing code—especially the boring parts early in your career.<br/><br/>So if you&#39;re new to this, turn off Github Copilot and struggle on some of it for longer than is comfortable. Learning takes place at the edge of comfort. 2025-07-08T13:01:11Z https://nostr.now/nevent1qqs2tp59rdymst8jpyfp39tmrye4wrlpp7p9j4e4af2mzmgk830gkyszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsct5hfs There&#39;s a pattern in software engineering: the best principles weren&#39;t planned—they were discovered.<br/><br/>Unix Philosophy (1978) emerged from Bell Labs after building Unix. Rob Pike&#39;s C Rules (1989) came after C proliferated. Both were written to explain what was already working.<br/><br/>&#34;The rules were discovered rather than declared.&#34;<br/><br/>This pattern repeats throughout software history... 1/3 2025-07-07T22:00:58Z https://nostr.now/nevent1qqs2gel9txxem8gcg72dz3syh95aclu26aafdph96u0t0y75mnz0jnczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs5ayss6 How do you primarily navigate large codebases? 2025-07-06T22:00:35Z https://nostr.now/nevent1qqswyf564ueep4acpq8hrxvyjv2sk56h7axmtep52a0kc58qx7yg0mczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrscuul6y &#34;We need bug-free code.&#34;<br/><br/>This sounds good. It&#39;s a trap.<br/><br/>Perfection isn&#39;t actionable. Better targets:<br/>• Rollback rate to 50% of current<br/>• Code coverage from unknown to 80%<br/>• X critical journeys pass Y times before release<br/><br/>Improvement against specific measures beats perfection in all things. 2025-06-28T01:00:09Z https://nostr.now/nevent1qqszpgqpjz8setur23y58kz32385gkyypyq9pk6n9cjfx2rl2q3udygzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsadtw9l The stories about 10x engineers never include:<br/>• Adding meaningful test coverage<br/>• Updating outdated README sections <br/>• Mentoring junior engineers<br/>• Listening to burned-out coworkers<br/>• Improving on-call runbooks<br/><br/>No glory in these things, but they&#39;re the ones you desperately want to keep.<br/><br/>They plant trees they don&#39;t expect to sit under. 2025-06-27T01:00:11Z https://nostr.now/nevent1qqsgfy72ntw6fh3kv7z58qcda0mt6sc0gqd0skg8s693l4lg7w89huszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs0lyyzr Goodhart&#39;s Law: &#34;When a measure becomes a target, it ceases to be a good measure.&#34;<br/><br/>You see this with:<br/>• Code coverage targets → meaningless tests<br/>• Story points → velocity games<br/>• Bug counts → closing tickets without fixing root causes<br/><br/>The best metrics are multiple, in tension with each other, and measure teams—not individuals. 2025-06-17T13:01:03Z https://nostr.now/nevent1qqs9n8hh762w6qu32as0vuudw66jvrfyftvr7zfh7vey06fz0ltjlcczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrshvkfrf AI isn&#39;t replacing new software engineers—it&#39;s hurting them.<br/><br/>New engineers use AI to skip the &#34;boring&#34; parts early in their career, but that&#39;s where you learn to think about problems.<br/><br/>It&#39;s like using a calculator for math homework. You get the answer but miss the learning.<br/><br/>Turn off Copilot sometimes. Struggle with the uncomfortable parts. Learning happens at the edge of comfort. 2025-06-12T01:00:43Z https://nostr.now/nevent1qqswghcsk4u4maemf4xsu294fncj4qsq3eyucsur9d805s5vux5j5kczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsdun4a9 I provide without any edits:<br/><br/>For Quartering large bodies of armed troops among us:<br/><br/>For protecting them, by a mock Trial, from punishment for any Murders which they should commit on the Inhabitants of these States:<br/><br/>For cutting off our Trade with all parts of the world:<br/><br/>For imposing Taxes on us without our Consent:<br/><br/>For depriving us in many cases, of the benefits of Trial by Jury: 2025-06-11T02:20:27Z https://nostr.now/nevent1qqspchy34yj6we6jllqxlq7py8zky8c6ehdvtjadr5jr8dtxuu9phxczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsaeepz4 Focus on effectiveness over efficiency. What&#39;s more important: solving a $1M problem at 90% efficiency or a $10K problem at 99.999% efficiency? Effectiveness differentiates great engineers. 2025-05-30T13:00:43Z https://nostr.now/nevent1qqsw6xnckt8tpelqhezdjtzmm8p96ca5m8y6kwdj2z5af7sdyfef49czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsfdhemp Setting clear expectations isn&#39;t mean - it&#39;s clear. Good bosses enforce realistic expectations. Bad bosses enforce none, suggesting the work doesn&#39;t matter. Worst bosses never set any at all. 2025-05-27T12:00:24Z https://nostr.now/nevent1qqsvzgdwxqefggfsfmf5l50jr6zazsg63suqlwj450pe72dq2466laczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs93ruqc &#34;Do as little testing as reasonable.&#34; Users will always find bugs - that&#39;s not enough signal to justify endless testing. Testing reduces risk but doesn&#39;t generate revenue. 2025-05-26T12:00:27Z https://nostr.now/nevent1qqsr535pktgln8ct8968aeh5qttt7s6fhknfaz3mxzm46kvrtq7qe9qzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs4vrcjy Variable naming conventions by programming language:<br/><br/>Python: the_thing<br/>JavaScript: theThing<br/>TypeScript: theThing: any<br/>Java: AbstractFactoryPatternSingletonInstanceManagerThing<br/>C: i<br/>Go: thing, err<br/>Rust: maybe_the_thing 2025-05-25T18:25:11Z https://nostr.now/nevent1qqsx0sldgppecmpr9xwgzgmx7aa47ny0ecvwurxy55jvsep4ku7khuczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsvhf87u I need some opinions, ya&#39;ll. <br/><br/>For distributing CLIs in a single binary in Linux/MacOS/Windows, what&#39;s your favorite language to do that?<br/><br/>Bonus question for the comments: what library is your go to in that language? 2025-05-24T13:01:29Z https://nostr.now/nevent1qqsz9q7l5m399599zumzv3y9kwmt3d966rm3mwjx20sg64jxzt0020czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrscngrqr The &#34;we need bug-free code&#34; mandate is a trap. It asks everyone to be perfect forever. Better targets: reduce rollback rate by 50%, increase code coverage to 80%, or define passing customer journeys. 2025-05-24T12:00:19Z https://nostr.now/nevent1qqs2hzwll7aq3xvpl07740lcke77egknjkwq5ajfjt3v4g83dqzzcpgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsr7uz04 Brook&#39;s Law remains true decades later: adding engineers to late projects makes them later. New team members require onboarding, communication channels increase exponentially, and integration points multiply. 2025-05-20T12:00:31Z https://nostr.now/nevent1qqswtgg3pernqck2vlaazp4sdu7a9u0mfvrqznr23248scln5q358fqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsy26avc Recognize the Blue Dot effect: as major problems are resolved, your threshold for what constitutes a &#34;problem&#34; shifts dramatically. Without calibration, teams spend disproportionate time on increasingly trivial issues. 2025-05-19T12:00:28Z https://nostr.now/nevent1qqs2hjyeuye3w88zut7nh83ljx2r037z0yc5wtu05eeaa0sw08x6a8szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsf7m2vm I&#39;m not anti-AI. I&#39;m just pro-thinking.<br/><br/>What do I mean? I mean that everyone I know who is REALLY into AI uses it for reading EVERYTHING longer than 1000 words.<br/><br/>This is the mental equivalent of eating only smoothies, because they go down easier.<br/><br/>I say this with love: reading is thinking.<br/><br/>Stop letting slop think for you. 2025-05-04T14:46:48Z https://nostr.now/nevent1qqswc7pprfy2stqrv7hxgljfthyx03drfcs28e9gh5fsmwsexy4gdhczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrszlugrv Would you rather permanently have?<br/><br/>(Boost for reach) 2025-03-12T22:33:44Z https://nostr.now/nevent1qqsdr7htj9jl3v7530m56jvrta9xrzt5ajpg4yxl4f7aj73hfrw9q5gzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsx4lx0m Things no one will mention at your funeral:<br/><br/>- participated in standups<br/>- increased to shareholder value<br/>- got in expense report on time all the time<br/>- had 90% code coverage on all PRs<br/>- clocked your time on time all the time<br/><br/>This is not an exhaustive list by any means. 2025-03-12T00:50:04Z https://nostr.now/nevent1qqs9pneh90uzyh50fftxg6xd4tlkkxg82hnwyxv6mnxa467hljt2fxqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs49vn0f In reply to <a href='/nevent1qqsyrr9w3r35t2zrh4xg9m9mvvzs6qpmysec2de7ymkwd5a7ydt37uq8kuu5u'>nevent1q…uu5u</a><br/>_________________________<br/><br/>The real constraints in modern software aren&#39;t usually about code generation. They&#39;re about:<br/><br/>- Understanding what users actually need<br/>- Intuiting bottlenecks on limited instrumentation <br/>- Identifying which components cause the most incidents<br/><br/>LLMs can&#39;t solve these problems because they require system-level thinking and real-world feedback.<br/><br/>The most effective engineers identify the right constraints to address. This skill becomes more valuable, not less, in an AI-assisted world.<br/><br/>2/2 2025-02-25T00:53:49Z https://nostr.now/nevent1qqsyrr9w3r35t2zrh4xg9m9mvvzs6qpmysec2de7ymkwd5a7ydt37uqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrshp29gw ChatGPT and Github Copilot can&#39;t help you with this problem.<br/><br/>To see it, let&#39;s consider three things that it will happily do if you ask:<br/><br/>- Generating perfect code for a feature that users don&#39;t need<br/>- Optimizing a service call that isn&#39;t actually your bottleneck<br/>- Creating more tests for stable components while unstable ones remain untested<br/><br/>What do these have in common? Solving the wrong constraint.<br/><br/>1/2 2025-02-25T00:53:04Z https://nostr.now/nevent1qqs2pp0c6pj2kyu8twyv03lfcucyuhzvy3df22uvm83gr2s6ksj6r5szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsvlamr7 Elon-sort: <br/>1. randomly eliminate half of the elements<br/>2. bring them back <br/>3. loop thru 1&amp;2 for a random number of times<br/>4. declare the array sorted without checking 2025-02-22T22:45:51Z https://nostr.now/nevent1qqs0mxxmkaqcz5vss2kw7tv458w8sr2t086k7jr46ff9tsaxn4zesuqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsacrewa TS with any is just normal JS.<br/> <img src="https://files.mastodon.social/media_attachments/files/114/045/003/831/693/518/original/16c09d6c50706b40.jpeg"> <br/> 2025-02-22T01:44:17Z https://nostr.now/nevent1qqszxspfzlcd3q587aw25qyn2dw6ehhahrrvttcxvr376e2653dt52czyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsjgsxez In reply to <a href='/nevent1qqszh0rjpegrq3pwvhsq2gswm7dmxy2f59kyaf7jwdwvdtg844jlj4ssd0dgp'>nevent1q…0dgp</a><br/>_________________________<br/><br/>What does emotional labor look like?<br/>- checking in with another manager who lashed out in a meeting yesterday<br/>- getting a &#34;I quit&#34; via text from a direct report and reaching out to say, &#34;hey, you want to just take a week off to think about it instead?&#34;<br/>- firing an under-performing person in such a way that they&#39;ll take your calls in 6 months<br/>- seeing the lights flicker on a remote call with your direct report in Kiev during March 2022 and not knowing if that&#39;s the last you&#39;ll ever see of them 2025-01-31T13:11:25Z https://nostr.now/nevent1qqszh0rjpegrq3pwvhsq2gswm7dmxy2f59kyaf7jwdwvdtg844jlj4szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsnj5a3a Engineering managers especially get this wrong.<br/><br/>&#34;We under value the &#39;emotional labor&#39; of being the boss.&#34; – Kim Scott, Radical Candor, p. 4 2025-01-31T13:09:41Z https://nostr.now/nevent1qqs0panjq5l256r9red0waeh7nn50u5j0vn22ddpsd4vr0a5nc9qnwszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsfjr3km I would tell you a joke about UDP...<br/><br/>but you might not get it. 2025-01-17T02:55:44Z https://nostr.now/nevent1qqsda7m7y0r9e9fjf3rhjpgetg3q84yqz59drxpgpxdu3zff9sqmgwszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs8cr6lr If the computer is a bicycle for the mind, then the smartphone is a hamster wheel. 🐹 2025-01-12T17:05:20Z https://nostr.now/nevent1qqsgs3sjm4zg7guhdya6q5lt4g0h2xmw2dw67rheper25mtvsvcyyqgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs8senr6 <img src="https://files.mastodon.social/media_attachments/files/113/812/368/870/061/645/original/9f91944132d59d31.png"> <br/> 2025-01-11T23:42:09Z https://nostr.now/nevent1qqsyf93mc8as36vsgeqnu9lyu79y5jgnjy93vxa78465vqypeuz0cqgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrslhneyz Any _good_ code generators out there?<br/><br/>I&#39;ve been working in Zed, using Claude 3.5 Sonnet. For completions, it&#39;s fine. No real complaints.<br/><br/>But when I try to do something larger (&gt;200 LOC), I&#39;m finding the output is often very bad. Doesn&#39;t matter the language.<br/><br/>It&#39;s likely that I&#39;m not able to express in a short block of text what I&#39;m trying to accomplish. So the small text prompt approach is likely the wrong UI. <br/><br/>But is there a better one? Has anyone seen it? 2024-12-15T15:11:48Z https://nostr.now/nevent1qqsvs3p2kuzlwvu8rht8mg697klqufy3hqzn0dz3p9dxsdfh2rrptzszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrstw883c Doing stuff feels good, and it&#39;s a trap.<br/><br/>You can spend forever coding or analyzing or architecting or designing or whatever doing your field does. <br/><br/>It feels good. It&#39;s often why you got into that.<br/><br/>Unless you&#39;re a musician or professional ball player, people aren&#39;t actually paying you for the doing. They&#39;re paying you to deliver something of value instead.<br/><br/>So enjoy what you&#39;re doing, and don&#39;t forget why you&#39;re doing it.<br/> <img src="https://files.mastodon.social/media_attachments/files/113/329/815/319/587/873/original/5e1637a8727a4df4.jpg"> <br/> 2024-10-18T18:22:57Z https://nostr.now/nevent1qqsq0ksq3g3zw4avyg9acs30ydvh0pysfq5st6rpv5qvqzve99wcm8szyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsk7jk3g In reply to <a href='/nevent1qqs2sx80aq7t6zznna9v2g3xczzj9hm8mt2xuq592zspg9ul42gspagu8g3gt'>nevent1q…g3gt</a><br/>_________________________<br/><br/>And not that it needs any additional proof, but I learned this most profoundly when I was talking to Guido van Rossum in person back like 7-8 years ago at a PyCon.<br/><br/>He mentioned off handedly, “Oh, I don’t get into packaging at all. I know very little about it, and I’m glad there others in our community who care to know about it.”<br/><br/>I know that people rag on Python’s ecosystem for dependency management as a bit scattershot, but it was my first good OSS ecosystem. (1/2) 2024-09-26T17:06:57Z https://nostr.now/nevent1qqs2sx80aq7t6zznna9v2g3xczzj9hm8mt2xuq592zspg9ul42gspagzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsw9zu7h Everyone starts as a bad programmer.<br/><br/>Everyone remains bad in the areas they don&#39;t learn about or practice. <br/><br/>That&#39;s OK.<br/><br/>Not every programmer knows everything.<br/><br/>Knowing your limits of your knowledge is as important as knowing things at all. 2024-09-26T16:45:00Z https://nostr.now/nevent1qqs8qdu97rxjpsjfudyujzcpy00qlkak06n7s6e2pd2ghrkjve8myvszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsch7yhx Do virtually all abstractions leak?<br/><br/>I recently watched a talk on how SQL leaks as an abstraction eventually. The speaker gave as an example the query planner. Spending enough time around SQL, you eventually need to run an explain to figure out what it&#39;s doing and why it&#39;s slow.<br/><br/>I think SQL in general does not leak too badly. So that it does eventually is fine.<br/><br/>What abstractions can you think of that encapsulate the complexity better? 2024-09-14T12:00:41Z https://nostr.now/nevent1qqsvfwswlvys80ua8mjtkety44lyvs6a37nhkapv58ta5ef8ltkcagczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrshh9qk6 If you have an idea and publish it on the internet, that counts as a ghost of done. 2024-09-03T18:30:57Z https://nostr.now/nevent1qqst6ze68uy68djcy6smfknzhggu7vvt58ejqc4ermspt4vdgv6xlrczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrslrnn99 Warning ⚠️ <br/><br/>Programming Dad joke<br/><video controls width="100%" class="max-h-[90vh] bg-neutral-300 dark:bg-zinc-700"><source src="https://files.mastodon.social/media_attachments/files/113/053/789/374/637/371/original/b7ef0e61fdee9f69.mp4"></video><br/> 2024-08-31T00:25:39Z https://nostr.now/nevent1qqsfkwkac56xve0xd0knwm5mftdtzn2zqxryrwzyfvwxyk4gmnad6agzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrs2nja4h In reply to <a href='/nevent1qqstww2htkeryark85r2rmxk6prp0h7ct2quq0swjj6dsyg36hg8mlq3fytf5'>nevent1q…ytf5</a><br/>_________________________<br/><br/>There exist things like:<br/>Python’s Pint: <a href="https://pint.readthedocs.io/en/stable/getting/tutorial.html">https://pint.readthedocs.io/en/stable/getting/tutorial.html</a><br/>Julia’s Unitful: <a href="https://juliahub.com/ui/Packages/General/Unitful">https://juliahub.com/ui/Packages/General/Unitful</a><br/><br/>And those are like….fine, but very much packages, not language fundamentals, and so cause execution time tradeoffs.<br/><br/>I’m talking about instead some kind of syntax that is a 0-cost abstraction. 2024-08-30T01:19:45Z https://nostr.now/nevent1qqstww2htkeryark85r2rmxk6prp0h7ct2quq0swjj6dsyg36hg8mlqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsual8jz In reply to <a href='/nevent1qqs8zu6w3lhv52jev82zefr066hzwvwl4gzk7lz93su9jn084gkldjce79ffe'>nevent1q…9ffe</a><br/>_________________________<br/><br/>The value of such a thing in my opinion would be for scientific computing where you could do all the unit analysis and verification that nothing silly happened, and you could check it as compile time. 2024-08-30T01:17:03Z https://nostr.now/nevent1qqs8zu6w3lhv52jev82zefr066hzwvwl4gzk7lz93su9jn084gkldjczyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsuty86p In reply to <a href='/nevent1qqs86xqts62ywaukatrqxu87n4vt8jtatt6e77wlnd0ny4xnwmt96vqy98t9s'>nevent1q…8t9s</a><br/>_________________________<br/><br/>Essentially, it would look like a language where a bunch of SI units are pushed into the language by default with a similar enum, but also offer a way for the user to extend the enum with other units. 2024-08-30T01:11:16Z https://nostr.now/nevent1qqs86xqts62ywaukatrqxu87n4vt8jtatt6e77wlnd0ny4xnwmt96vqzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsxap9pg In reply to <a href='/nevent1qqsdcdz44deqrsc72yqzvexlgsnplaj0rle95un4gln8h8a9g5vlrps0qk7q4'>nevent1q…k7q4</a><br/>_________________________<br/><br/>I’ve often wondered what it would be like to push the unit into some programming language primitive.<br/><br/>In particular, I’ve been inspired by Rust’s result type in the past: <a href="https://doc.rust-lang.org/std/result/">https://doc.rust-lang.org/std/result/</a> 2024-08-30T01:10:43Z https://nostr.now/nevent1qqsdcdz44deqrsc72yqzvexlgsnplaj0rle95un4gln8h8a9g5vlrpszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrse0hl8z In reply to <a href='/nevent1qqsrtpzrafed98gcvnw2pwjns76q6w7ww40yysd9k5xj7guzf0y7qls3ssssq'>nevent1q…sssq</a><br/>_________________________<br/><br/>However, many times programmers just forget or consider context clear enough, and thus we get variable like:<br/><br/>num, count, or i/j/k 2024-08-30T01:06:36Z https://nostr.now/nevent1qqsrtpzrafed98gcvnw2pwjns76q6w7ww40yysd9k5xj7guzf0y7qlszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsej04r0 Ever noticed how programming languages are &#34;unitless&#34;?<br/><br/>Like for instance,<br/><br/>let apple_count: number = 1;<br/><br/>We generally encode the unit information 𝘪𝘯𝘵𝘰 the variable name. 2024-08-30T01:04:38Z https://nostr.now/nevent1qqsr5meknna3u7n3nyjmxm54lqev0nujwemsegwclsexd4lzfkmvrkgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsjqnce8 LLMs are exo-skeletons for devs, not skeletons. <br/><br/>You can&#39;t build an engineer out of one, but you sure can empower one. 2024-08-29T15:11:33Z https://nostr.now/nevent1qqs0q9p4v3afuafk9mgdjleggcurxcnu38q54ss8ukrcxq6ymuczcmgzyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsppt58p Makes you wonder what all that RTO was all about. 🤔<br/><br/>src: <a href="https://www.instagram.com/p/Ch3ca-bL1te/">https://www.instagram.com/p/Ch3ca-bL1te/</a><br/> <img src="https://files.mastodon.social/media_attachments/files/113/028/211/228/791/521/original/731b620860b5f586.png"> <br/> 2024-08-26T12:00:33Z https://nostr.now/nevent1qqs2r7egtu29k6ra0j5tljlyx05xq73j0gd49xhk8nk7azws6ej66pszyrnu5wx3azwawpzetpmrtn80jafltxq5cy4cfmk34q3xxu8y05mrsv28ttr People over value efficiency and under value 𝘯𝘰𝘵 doing stuff: when it takes you 0 hours to do something 0 times, since you said “No”, it has an immediate payback period! 2024-08-15T11:55:35Z