Filter
Exclude
Time range
-
Near
What we do as engineers is changing. Design the right tools for that. news.ycombinator.com/item?id… 1. shmel - Uses plan mode as a default for non-trivial work. States that plan mode helps tremendously, and he only skips it for small things. He also emphasizes insisting on a strict verification suite alongside planning. 2. pgt - Follows a structured loop: spec → plan → critique → improve plan → implement plan. He later condensed this into two prompts: write the plan in Plan Mode, then exit Plan Mode and run a critique/improve loop before implementing. 3. CharlesW - Builds on pgt's flow by adding a code review step at the end, noting that reviewing an implementation with the same model you used to write it works extremely well, especially with the effort knob cranked to max. 4. spicyusername - Describes his workflow shift: throw as much context into Opus as possible about what I want in plan mode, then spend up to 60 minutes refining the plan before handing off implementation. This replaced his old habit of spending 0–360 minutes reading through code before writing anything. 5. x3n0ph3n3 - Happy with a Plan → Revise → Build → Correct flow in Cursor. Also shares plan documents with teammates: he posts his Cursor Plan document and Transcript in PRs so others can learn from it. 6. crsl - Says what works is iterating on a plan then implementing and reviewing, and that the more constraints he puts into the plan the better. 7. cloud8421 - Reports that incremental changes tend to be well done and mostly one-shotted provided he uses plan mode first, and that larger changes require careful planning with split phases. 8. turlockmike - Advocates spending time upfront reviewing the plan, and says if you clearly communicate intent and desired outcomes rather than just saying "do X," the quality increases significantly. 9. wombat-man - Found that once the plan is set, using the agentic coder to create smaller CLs (changelists) has been the best approach. Emphasizes you don't want to generate code faster than you and your reviewers can comprehend it. 10. konaraddi - Notes a downside: it often requires a lot of upfront planning and feedback loop setup, to the extent that he sometimes wonders if doing it himself would have been faster. 11. mentalgear - Warns against planning with the LLM too early. He's moved away from using an LLM before having figured out the specifications, because otherwise it's risky to go down a wrong rabbit hole the LLM lures you into with its positive engagement feedback. 12. ruduhudi - Highlights a limitation of plan-driven workflows: Claude performs well on tasks where you can write a precise spec, but changing code manually is more efficient for discovery work than dealing with large chunks of constantly changing AI-generated code when you don't yet know how the feature should look. 13. seanmcdirmid - Transitioning to AI-driven development, which he describes as focusing on your specification since you are handing development off to the AI afterwards. 14. anyonecancode - Uses Plan mode for the initial exploration phase, prompting with detailed descriptions and asking the model to review the codebase for similar patterns before proposing a plan. Notes that even a few months ago on earlier, not-as-performant models, he was still finding value with this approach. 15. alexmuro - Spends more than half his time with AI adding context, plans, and documentation to the repo. Views deciding what to work on next as the hardest part — figuring out what to do next that creates value for users is the hard part, whether delegating to humans or AI. 16. parasti - Describes a spec-driven multi-agent pattern: one session writes an API and its specification, another implements the client from that spec. As edge cases emerged, he had the first agent document them in the spec and the second agent read the updated spec, which worked much better than ad-hoc back and forth.

2
1
9
530
How to code your own indicators with ai's help , popularly knowns as vibe coding . Let me show you how i coded this 1. First i i had typed the same thing i.e " "Triple Tight" indicator .It appears when the combined 3 days range of the last 3 candle is < ATR14 Basically highlights Mega Tightness. Code this in pinescript version 5 and re check the code for any issues 2. checked and found out that it is not plotting anything on chart 3. Corrected on how we are basically defining this combined 3 days range as "combined three days range here means that we take the range of 3 candles max high and max low and then check if it is < atr 14" and voila !! Yes its that easy to code #anyonecancode #Tradingview #vibecoding @SakatasHomma @finallynitin x.com/SakatasHomma/status/19…
Yellow Star is my proprietary "Triple Tight" indicator. It appears when the combined 3 days range of the last 3 candle is < ATR14 Basically highlights Mega Tightness. I explain it in more detail in this video: youtube.com/watch?v=-61yuPMI… This is a Tightness Trading system. Some 30 odd names today in the Market.
4
14
102
14,826
anyOneCanCode redd.it/1it5cw4
2
4
47
7,861
Loved getting to work with our SLC students today and exposing them to ozobots again! Thanks Mr Barfield for letting us join you. They seemed to enjoy it, and I know I sure did! 😍@EHSFrisco @Ozobot #futureready #HourofCode #anyonecancode
7
105
25 Nov 2024
I'm back with the #javascript tips :) Tip 8 for complete beginners Use Object Property Shorthand When defining objects, if the property name is the same as the variable name, you can use shorthand: How does it Work: When you write name in the object literal, JavaScript automatically understands that you mean name: name. This shorthand syntax applies to any variable you wish to use as a key-value pair in an object where the key and the variable name match. #codingcommunity #anyonecancode
16 Nov 2024
#javascript Tip 7 for beginners Do you know about the  Short Circuit Evaluation in JS using && operator ?  It means that the && (logical AND) operator will only evaluate the right-hand operand if the left-hand operand is truthy. If the left operand is falsy, the right operand isn't evaluated at all. Practical Use: The && operator can be used to conditionally execute code based on the truthiness of a value. Here’s how it works: In the first case, user is null, which is falsy. Therefore, console.log(user.name) is not executed because of short-circuiting. The expression user &&. console.log(user.name) evaluates to null. In the second case, user is an object with a name property, which is truthy. Therefore, the expression evaluates to console.log(user.name), which executes and logs "Alice" to the console.
5
195
🚛 #DigiTruckEthiopia is back on the road, delivering hands-on digital literacy and coding skills to students across #Ethiopia! Empowering the next generation through tech. 💻✨ Click to learn more: bit.ly/3A0aMRQ #DigiTruck #AnyoneCanCode #iCog #TECH4ALL #DigitalSkills
4
10
123
735,053
Coding on Swift Playground is so much fun! #AnyoneCanCode #NewProfilePic
1
2
145
Bringing classrooms to doorsteps! For students traveling miles for education, DigiTruck Ethiopia is on a mission to deliver quality learning directly to them. #Digitruckethiopia #Tech4all #Anyonecancode #Youthempowerment
1
1
7
5,085
S3 and S4 @PrincesTrust Achieve Clubs having a great time “Bridging the App”. Honing our coding skills and creating and presenting our own apps. The pizza buffet was very welcome after all our hard work! #Ambition #AnyoneCanCode
1
4
4,534
We did it! We completed our mission to bring digital skills and education to the students of Hawassa. Learning digital skills with #digitrucket #anyonecancode #TECH4ALL #fdremoe #steameducation #ethiopianyouth #tech #ethiotech @iCog_ACC @Huawei @fdremoe
2
7
8,895
Reason No. 12,082,023 to love 8th grade…. Having fun with Hour of Code #hourofcode #anyonecancode
2
61
#CSEdWeek kicks off today & @srmsct will be celebrating #HourofCode all month. Our SRMS Morning Show will feature ways to inspire K-12 students to learn about computer science and to embark on their own coding journey. @RidgefieldPS @codeorg #anyonecancode #edtech #futurecoders
3
95
The #hourofcode is here in the Lebanon School District!! We had our kick-off Zoom this morning and we were excited that so many classrooms joined us this morning. @codeorg #anyonecancode
1
3
19
1,476
Loving Science Week! Great to be back in St Andrew's College Junior School on this windy day to deliver workshops to their wonderful students! 👩‍💻👨‍💻#scienceweek #anyonecancode @SACjuniorschool
2
207
🔥 New feature in Codelita! 🔥 Now when you solve a problem, you have the option of "View Codelita's Solution." You can also go back and check our solutions retrospectively! We hope it will enhance your learning and boost your confidence in coding. 😍 #Codelita #AnyoneCanCode
1
36
17,658
Congratulations to Ladan for spending 20 hours on @theCodelita and finishing her first course with us! Well done. 😍🥳👏 #Codelita #AnyoneCanCode linkedin.com/posts/eyedean_i…
24
2,371
A wonderful Halloween Camp this week @MERITS_Hub Monster Mash, Bat-hunter, Apple pong, Star Catcher and Boom Boom Race are just some of the cool projects these fab kids designed and coded. #anyonecancode
2
157
23 Oct 2023
Not 10, not 1000, not 10,000, but 100,000 codes ran and submitted in Codelita! 🥳 Next milestone: 1,000,000 codes! 😇 #Codelita #AnyoneCanCode
14
1,930
23 Oct 2023
We just surpassed 100,000 code submissions on Codelita! 😍 Join us now: cod.la #100KMilestone #Codelita #AnyoneCanCode
1
17
1,724
In just four regional cities of Ethiopia, DigiTruck Ethiopia has been able to narrow the digital divide for hundreds of young people. #digitrucket #anyonecancode #TECH4ALL #fdremoe #steameducation #youthempowerment #ethiopianyouth #tech #ethiotech @iCog_ACC @Huawei @fdremoe
39
17
36,002