#devex — Public Fediverse posts

Реальный DX: как измерить опыт разработчика и не соврать самому себе

В прошлый раз я писал про то, что сообщение об ошибке читает уставший человек в два часа ночи. Там была одна простая мысль: ошибка должна сказать, что случилось, почему и что делать дальше. Переписали голое invalid_request в человеческий ответ, и стало лучше. Вот только «стало лучше» — это ощущение. А ощущение нельзя положить в роадмап (или защитить перед руководством), чтобы сравнить через квартал. Улучшить то, что ты не измеряешь, нельзя: можно только верить, что улучшил. Давайте сегодня разберёмся, как DX из набора интуиций превратился в измеримую дисциплину: с исследованиями, метриками и цифрами. И как этим пользоваться, не скатываясь в дашборд ради дашборда.

https://habr.com/ru/articles/1048262/

#dx #api #developer_experience #опыт_разработчика #devex #space #dora #DX_core_4 #ttfhw #когнитивная_нагрузка

Реальный DX: как измерить опыт разработчика и не соврать самому себе

В прошлый раз я писал про то, что сообщение об ошибке читает уставший человек в два часа ночи. Там была одна простая мысль: ошибка должна сказать, что случилось, почему и что делать дальше. Переписали голое invalid_request в человеческий ответ, и стало лучше. Вот только «стало лучше» — это ощущение. А ощущение нельзя положить в роадмап (или защитить перед руководством), чтобы сравнить через квартал. Улучшить то, что ты не измеряешь, нельзя: можно только верить, что улучшил. Давайте сегодня разберёмся, как DX из набора интуиций превратился в измеримую дисциплину: с исследованиями, метриками и цифрами. И как этим пользоваться, не скатываясь в дашборд ради дашборда.

https://habr.com/ru/articles/1048262/

#dx #api #developer_experience #опыт_разработчика #devex #space #dora #DX_core_4 #ttfhw #когнитивная_нагрузка

Реальный DX: как измерить опыт разработчика и не соврать самому себе

В прошлый раз я писал про то, что сообщение об ошибке читает уставший человек в два часа ночи. Там была одна простая мысль: ошибка должна сказать, что случилось, почему и что делать дальше. Переписали голое invalid_request в человеческий ответ, и стало лучше. Вот только «стало лучше» — это ощущение. А ощущение нельзя положить в роадмап (или защитить перед руководством), чтобы сравнить через квартал. Улучшить то, что ты не измеряешь, нельзя: можно только верить, что улучшил. Давайте сегодня разберёмся, как DX из набора интуиций превратился в измеримую дисциплину: с исследованиями, метриками и цифрами. И как этим пользоваться, не скатываясь в дашборд ради дашборда.

https://habr.com/ru/articles/1048262/

#dx #api #developer_experience #опыт_разработчика #devex #space #dora #DX_core_4 #ttfhw #когнитивная_нагрузка

Reading about special relativity felt like understanding it. Doing the calculations in university proved it wasn't the same thing.

AI coding gives me a similar uneasy feeling. I can follow the logic and reason about the system, but knowing how something works is not the same as being able to apply it. The applied judgment can quietly move out of the developer's hands.

What bothers me is this: if you need AI to understand your own codebase, that understanding is no longer just assisted by a tool. It is delegated to a vendor.

#AI #DevEx

Reading about special relativity felt like understanding it. Doing the calculations in university proved it wasn't the same thing.

AI coding gives me a similar uneasy feeling. I can follow the logic and reason about the system, but knowing how something works is not the same as being able to apply it. The applied judgment can quietly move out of the developer's hands.

What bothers me is this: if you need AI to understand your own codebase, that understanding is no longer just assisted by a tool. It is delegated to a vendor.

#AI #DevEx

Reading about special relativity felt like understanding it. Doing the calculations in university proved it wasn't the same thing.

AI coding gives me a similar uneasy feeling. I can follow the logic and reason about the system, but knowing how something works is not the same as being able to apply it. The applied judgment can quietly move out of the developer's hands.

What bothers me is this: if you need AI to understand your own codebase, that understanding is no longer just assisted by a tool. It is delegated to a vendor.

#AI #DevEx

Reading about special relativity felt like understanding it. Doing the calculations in university proved it wasn't the same thing.

AI coding gives me a similar uneasy feeling. I can follow the logic and reason about the system, but knowing how something works is not the same as being able to apply it. The applied judgment can quietly move out of the developer's hands.

What bothers me is this: if you need AI to understand your own codebase, that understanding is no longer just assisted by a tool. It is delegated to a vendor.

#AI #DevEx

RE: https://mastodon.social/@heplon/116687109227735875

After years of watching great engineers lose entire afternoons to pipeline debugging, CrashLoopBackoffs and OOM kills, my co-founder Simon and I have decided to do something about it.

Today we are launching Heplon!

#PlatformEngineering #DeveloperExperience #DevEx #InternalDeveloperPlatform

Rails 8.1 adds Local CI: a config/ci.rb DSL plus bin/ci runner that executes the same checks on your laptop and in GitHub Actions.

New apps come with #Brakeman, bundler-audit, #Rubocop, and system tests preconfigured.

https://go.fastruby.io/djq

#RubyOnRails #Rails81 #DevEx

Rails 8.1 adds Local CI: a config/ci.rb DSL plus bin/ci runner that executes the same checks on your laptop and in GitHub Actions.

New apps come with #Brakeman, bundler-audit, #Rubocop, and system tests preconfigured.

https://go.fastruby.io/djq

#RubyOnRails #Rails81 #DevEx

I wrote my own coding agent cost tracker and noticed that my numbers did not line up with codexbar by Peter Steinberger.

so I did the annoying thing and dug into the details, assuming I had probably made the mistake.

the mismatch still held.

with the public “$1.3M in tokens in 30 days” context, I keep wondering:

if the cost math of his software is off, how much of the software produced at that scale is actually right?

#DevEx #Programming #AI

I wrote my own coding agent cost tracker and noticed that my numbers did not line up with codexbar by Peter Steinberger.

so I did the annoying thing and dug into the details, assuming I had probably made the mistake.

the mismatch still held.

with the public “$1.3M in tokens in 30 days” context, I keep wondering:

if the cost math of his software is off, how much of the software produced at that scale is actually right?

#DevEx #Programming #AI

I wrote my own coding agent cost tracker and noticed that my numbers did not line up with codexbar by Peter Steinberger.

so I did the annoying thing and dug into the details, assuming I had probably made the mistake.

the mismatch still held.

with the public “$1.3M in tokens in 30 days” context, I keep wondering:

if the cost math of his software is off, how much of the software produced at that scale is actually right?

#DevEx #Programming #AI

I wrote my own coding agent cost tracker and noticed that my numbers did not line up with codexbar by Peter Steinberger.

so I did the annoying thing and dug into the details, assuming I had probably made the mistake.

the mismatch still held.

with the public “$1.3M in tokens in 30 days” context, I keep wondering:

if the cost math of his software is off, how much of the software produced at that scale is actually right?

#DevEx #Programming #AI

I wrote my own coding agent cost tracker and noticed that my numbers did not line up with codexbar by Peter Steinberger.

so I did the annoying thing and dug into the details, assuming I had probably made the mistake.

the mismatch still held.

with the public “$1.3M in tokens in 30 days” context, I keep wondering:

if the cost math of his software is off, how much of the software produced at that scale is actually right?

#DevEx #Programming #AI

From engineering as we know it ➔ AI-native engineering.

Inside #Meta’s “AI for Productivity” (AI4P) initiative: what they tried, what worked, what still needs proof, and what they’ve ruled out (for now).

Watch #InfoQ video for more insights: https://bit.ly/4tRTzQA

#AI #SoftwareEngineering #EngineeringLeadership #AgenticAI #DevEx

From engineering as we know it ➔ AI-native engineering.

Inside #Meta’s “AI for Productivity” (AI4P) initiative: what they tried, what worked, what still needs proof, and what they’ve ruled out (for now).

Watch #InfoQ video for more insights: https://bit.ly/4tRTzQA

#AI #SoftwareEngineering #EngineeringLeadership #AgenticAI #DevEx

From engineering as we know it ➔ AI-native engineering.

Inside #Meta’s “AI for Productivity” (AI4P) initiative: what they tried, what worked, what still needs proof, and what they’ve ruled out (for now).

Watch #InfoQ video for more insights: https://bit.ly/4tRTzQA

#AI #SoftwareEngineering #EngineeringLeadership #AgenticAI #DevEx

From engineering as we know it ➔ AI-native engineering.

Inside #Meta’s “AI for Productivity” (AI4P) initiative: what they tried, what worked, what still needs proof, and what they’ve ruled out (for now).

Watch #InfoQ video for more insights: https://bit.ly/4tRTzQA

#AI #SoftwareEngineering #EngineeringLeadership #AgenticAI #DevEx

From engineering as we know it ➔ AI-native engineering.

Inside #Meta’s “AI for Productivity” (AI4P) initiative: what they tried, what worked, what still needs proof, and what they’ve ruled out (for now).

Watch #InfoQ video for more insights: https://bit.ly/4tRTzQA

#AI #SoftwareEngineering #EngineeringLeadership #AgenticAI #DevEx

#PlatformEngineering succeeds when reliability and ergonomics reinforce each other - not compete.

Pratik Agarwal explores 3 foundational pillars:
1️⃣ Automated reliability
2️⃣ Developer ergonomics
3️⃣ Operator ergonomics

The result❓ Together, they create a Virtuous Cycle! 🔄

Read the #InfoQ article for more insights: https://bit.ly/4ew5WxK

#DevOps #Observability #SRE #DevEx

#PlatformEngineering succeeds when reliability and ergonomics reinforce each other - not compete.

Pratik Agarwal explores 3 foundational pillars:
1️⃣ Automated reliability
2️⃣ Developer ergonomics
3️⃣ Operator ergonomics

The result❓ Together, they create a Virtuous Cycle! 🔄

Read the #InfoQ article for more insights: https://bit.ly/4ew5WxK

#DevOps #Observability #SRE #DevEx

#PlatformEngineering succeeds when reliability and ergonomics reinforce each other - not compete.

Pratik Agarwal explores 3 foundational pillars:
1️⃣ Automated reliability
2️⃣ Developer ergonomics
3️⃣ Operator ergonomics

The result❓ Together, they create a Virtuous Cycle! 🔄

Read the #InfoQ article for more insights: https://bit.ly/4ew5WxK

#DevOps #Observability #SRE #DevEx

#PlatformEngineering succeeds when reliability and ergonomics reinforce each other - not compete.

Pratik Agarwal explores 3 foundational pillars:
1️⃣ Automated reliability
2️⃣ Developer ergonomics
3️⃣ Operator ergonomics

The result❓ Together, they create a Virtuous Cycle! 🔄

Read the #InfoQ article for more insights: https://bit.ly/4ew5WxK

#DevOps #Observability #SRE #DevEx

#PlatformEngineering succeeds when reliability and ergonomics reinforce each other - not compete.

Pratik Agarwal explores 3 foundational pillars:
1️⃣ Automated reliability
2️⃣ Developer ergonomics
3️⃣ Operator ergonomics

The result❓ Together, they create a Virtuous Cycle! 🔄

Read the #InfoQ article for more insights: https://bit.ly/4ew5WxK

#DevOps #Observability #SRE #DevEx

DORA-метрики: как собирать, интерпретировать и не переусердствовать, часть 2

В первой части мы разобрали, как устроены DORA-метрики и что стоит за каждым из пяти показателей. Сложнее другое: одни используют их как инструмент улучшения процессов, другие – как универсальную шкалу зрелости. Разбираемся, почему контекст здесь важнее любого бенчмарка – и с чего начать команде, которая хочет считать метрики осмысленно.

https://habr.com/ru/companies/simpleone/articles/1026602/

#DORA #devops #метрики_разработки #deployment_frequency #cicd #software_delivery #бенчмарки #управление_разработкой #itинфраструктура #devex

DORA-метрики: как собирать, интерпретировать и не переусердствовать, часть 2

В первой части мы разобрали, как устроены DORA-метрики и что стоит за каждым из пяти показателей. Сложнее другое: одни используют их как инструмент улучшения процессов, другие – как универсальную шкалу зрелости. Разбираемся, почему контекст здесь важнее любого бенчмарка – и с чего начать команде, которая хочет считать метрики осмысленно.

https://habr.com/ru/companies/simpleone/articles/1026602/

#DORA #devops #метрики_разработки #deployment_frequency #cicd #software_delivery #бенчмарки #управление_разработкой #itинфраструктура #devex

DORA-метрики: как собирать, интерпретировать и не переусердствовать, часть 2

В первой части мы разобрали, как устроены DORA-метрики и что стоит за каждым из пяти показателей. Сложнее другое: одни используют их как инструмент улучшения процессов, другие – как универсальную шкалу зрелости. Разбираемся, почему контекст здесь важнее любого бенчмарка – и с чего начать команде, которая хочет считать метрики осмысленно.

https://habr.com/ru/companies/simpleone/articles/1026602/

#DORA #devops #метрики_разработки #deployment_frequency #cicd #software_delivery #бенчмарки #управление_разработкой #itинфраструктура #devex

Anysphere released #Cursor 3 - a redesigned interface built from scratch that shifts the primary model from file editing ⇨ managing parallel coding agents.

The new workspace supports local-to-cloud agent handoff, multi-repo parallel execution, and a plugin marketplace.

Details here ⇨ https://bit.ly/4ezYNfH

#InfoQ #AIdevelopment #LLMs #DevEx

Anysphere released #Cursor 3 - a redesigned interface built from scratch that shifts the primary model from file editing ⇨ managing parallel coding agents.

The new workspace supports local-to-cloud agent handoff, multi-repo parallel execution, and a plugin marketplace.

Details here ⇨ https://bit.ly/4ezYNfH

#InfoQ #AIdevelopment #LLMs #DevEx

Anysphere released #Cursor 3 - a redesigned interface built from scratch that shifts the primary model from file editing ⇨ managing parallel coding agents.

The new workspace supports local-to-cloud agent handoff, multi-repo parallel execution, and a plugin marketplace.

Details here ⇨ https://bit.ly/4ezYNfH

#InfoQ #AIdevelopment #LLMs #DevEx

Anysphere released #Cursor 3 - a redesigned interface built from scratch that shifts the primary model from file editing ⇨ managing parallel coding agents.

The new workspace supports local-to-cloud agent handoff, multi-repo parallel execution, and a plugin marketplace.

Details here ⇨ https://bit.ly/4ezYNfH

#InfoQ #AIdevelopment #LLMs #DevEx

Anysphere released #Cursor 3 - a redesigned interface built from scratch that shifts the primary model from file editing ⇨ managing parallel coding agents.

The new workspace supports local-to-cloud agent handoff, multi-repo parallel execution, and a plugin marketplace.

Details here ⇨ https://bit.ly/4ezYNfH

#InfoQ #AIdevelopment #LLMs #DevEx

I'm working on an educational app for #amateurradio. In it I have visualizations for different factors that affect radio signal propogation.

My AI agents hit a quality plateau. Prompt tweaks weren't helping.

The problem? My agent was anchored to MY understanding of the domain, reflecting my blind spots back at me. 🤦‍♂️

Here's how I got through it:
https://philjohnstonii.com/blog/fresh-eyes-how-i-got-past-my-agents-quality-ceiling

#AI #AgenticCoding #LLM #DevEx #nvis #hamradio #hf #pota

I'm working on an educational app for #amateurradio. In it I have visualizations for different factors that affect radio signal propogation.

My AI agents hit a quality plateau. Prompt tweaks weren't helping.

The problem? My agent was anchored to MY understanding of the domain, reflecting my blind spots back at me. 🤦‍♂️

Here's how I got through it:
https://philjohnstonii.com/blog/fresh-eyes-how-i-got-past-my-agents-quality-ceiling

#AI #AgenticCoding #LLM #DevEx #nvis #hamradio #hf #pota

I'm working on an educational app for #amateurradio. In it I have visualizations for different factors that affect radio signal propogation.

My AI agents hit a quality plateau. Prompt tweaks weren't helping.

The problem? My agent was anchored to MY understanding of the domain, reflecting my blind spots back at me. 🤦‍♂️

Here's how I got through it:
https://philjohnstonii.com/blog/fresh-eyes-how-i-got-past-my-agents-quality-ceiling

#AI #AgenticCoding #LLM #DevEx #nvis #hamradio #hf #pota

I'm working on an educational app for #amateurradio. In it I have visualizations for different factors that affect radio signal propogation.

My AI agents hit a quality plateau. Prompt tweaks weren't helping.

The problem? My agent was anchored to MY understanding of the domain, reflecting my blind spots back at me. 🤦‍♂️

Here's how I got through it:
https://philjohnstonii.com/blog/fresh-eyes-how-i-got-past-my-agents-quality-ceiling

#AI #AgenticCoding #LLM #DevEx #nvis #hamradio #hf #pota

I'm working on an educational app for #amateurradio. In it I have visualizations for different factors that affect radio signal propogation.

My AI agents hit a quality plateau. Prompt tweaks weren't helping.

The problem? My agent was anchored to MY understanding of the domain, reflecting my blind spots back at me. 🤦‍♂️

Here's how I got through it:
https://philjohnstonii.com/blog/fresh-eyes-how-i-got-past-my-agents-quality-ceiling

#AI #AgenticCoding #LLM #DevEx #nvis #hamradio #hf #pota

AI tools are not “bad”.
But they often give bad results.

Why? Because they don’t understand your codebase.
And your prompts are missing context.

This is the real gap most Java developers hit right now.

I break it down here:
https://www.the-main-thread.com/p/java-developers-ai-tools-bad-results-context-prompting

#Java #AI #Quarkus #DevEx #LLM #SoftwareEngineering

AI tools are not “bad”.
But they often give bad results.

Why? Because they don’t understand your codebase.
And your prompts are missing context.

This is the real gap most Java developers hit right now.

I break it down here:
https://www.the-main-thread.com/p/java-developers-ai-tools-bad-results-context-prompting

#Java #AI #Quarkus #DevEx #LLM #SoftwareEngineering

AI tools are not “bad”.
But they often give bad results.

Why? Because they don’t understand your codebase.
And your prompts are missing context.

This is the real gap most Java developers hit right now.

I break it down here:
https://www.the-main-thread.com/p/java-developers-ai-tools-bad-results-context-prompting

#Java #AI #Quarkus #DevEx #LLM #SoftwareEngineering

AI tools are not “bad”.
But they often give bad results.

Why? Because they don’t understand your codebase.
And your prompts are missing context.

This is the real gap most Java developers hit right now.

I break it down here:
https://www.the-main-thread.com/p/java-developers-ai-tools-bad-results-context-prompting

#Java #AI #Quarkus #DevEx #LLM #SoftwareEngineering

AI tools are not “bad”.
But they often give bad results.

Why? Because they don’t understand your codebase.
And your prompts are missing context.

This is the real gap most Java developers hit right now.

I break it down here:
https://www.the-main-thread.com/p/java-developers-ai-tools-bad-results-context-prompting

#Java #AI #Quarkus #DevEx #LLM #SoftwareEngineering

Rails logs in development: beautiful stack traces, everything in your terminal, life is good.

Rails logs in production three weeks later: zgrep through 30 compressed files and pray.

I wrote about why this happens and how to fix it.

https://rubystacknews.com/2026/03/31/rails-logs-a-blessing-in-development-a-headache-in-production/

#Rails #RubyOnRails #Ruby #Logging #DevEx #WebDev

#DeveloperProductivity isn’t just about faster builds or better tools - it’s about creating an environment where developers can do their best work.

In this #InfoQ video, Nicole Forsgren shares a practical blueprint to:
• Identify & remove the hidden friction slowing your team down
• Uncover obstacles
• Win stakeholder buy-in
• Scale DevEx improvements across your organization

🎬 Watch now | 📄 #transcript included: https://bit.ly/4tivSAP

#DevEx #BestPractices #Agile

#DeveloperProductivity isn’t just about faster builds or better tools - it’s about creating an environment where developers can do their best work.

In this #InfoQ video, Nicole Forsgren shares a practical blueprint to:
• Identify & remove the hidden friction slowing your team down
• Uncover obstacles
• Win stakeholder buy-in
• Scale DevEx improvements across your organization

🎬 Watch now | 📄 #transcript included: https://bit.ly/4tivSAP

#DevEx #BestPractices #Agile

#DeveloperProductivity isn’t just about faster builds or better tools - it’s about creating an environment where developers can do their best work.

In this #InfoQ video, Nicole Forsgren shares a practical blueprint to:
• Identify & remove the hidden friction slowing your team down
• Uncover obstacles
• Win stakeholder buy-in
• Scale DevEx improvements across your organization

🎬 Watch now | 📄 #transcript included: https://bit.ly/4tivSAP

#DevEx #BestPractices #Agile

#DeveloperProductivity isn’t just about faster builds or better tools - it’s about creating an environment where developers can do their best work.

In this #InfoQ video, Nicole Forsgren shares a practical blueprint to:
• Identify & remove the hidden friction slowing your team down
• Uncover obstacles
• Win stakeholder buy-in
• Scale DevEx improvements across your organization

🎬 Watch now | 📄 #transcript included: https://bit.ly/4tivSAP

#DevEx #BestPractices #Agile

#DeveloperProductivity isn’t just about faster builds or better tools - it’s about creating an environment where developers can do their best work.

In this #InfoQ video, Nicole Forsgren shares a practical blueprint to:
• Identify & remove the hidden friction slowing your team down
• Uncover obstacles
• Win stakeholder buy-in
• Scale DevEx improvements across your organization

🎬 Watch now | 📄 #transcript included: https://bit.ly/4tivSAP

#DevEx #BestPractices #Agile