AI Native Dev
@ainativedev.io
Your weekly digest of AI software trends, expert insights, & handpicked content and events, delivered straight to your inbox.
https://ainativedev.io/newsletter
https://ainativedev.io/newsletter
AI hates legacy code.
AI coding agents feel like magic... right up until they collide with production code.
For teams maintaining legacy systems, these agents often hallucinate APIs, run off on tangents, and shatter trust faster than an unreviewed hotfix at 5pm.
🧵 1/3
AI coding agents feel like magic... right up until they collide with production code.
For teams maintaining legacy systems, these agents often hallucinate APIs, run off on tangents, and shatter trust faster than an unreviewed hotfix at 5pm.
🧵 1/3
November 12, 2025 at 12:30 PM
AI hates legacy code.
AI coding agents feel like magic... right up until they collide with production code.
For teams maintaining legacy systems, these agents often hallucinate APIs, run off on tangents, and shatter trust faster than an unreviewed hotfix at 5pm.
🧵 1/3
AI coding agents feel like magic... right up until they collide with production code.
For teams maintaining legacy systems, these agents often hallucinate APIs, run off on tangents, and shatter trust faster than an unreviewed hotfix at 5pm.
🧵 1/3
What is happening in the world of AI native development today?
These are builders who've already solved the problems you're probably hitting right now.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
Use the discount code AIND-X-50 to get 50% off your tickets.
These are builders who've already solved the problems you're probably hitting right now.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
Use the discount code AIND-X-50 to get 50% off your tickets.
November 11, 2025 at 8:51 AM
What is happening in the world of AI native development today?
These are builders who've already solved the problems you're probably hitting right now.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
Use the discount code AIND-X-50 to get 50% off your tickets.
These are builders who've already solved the problems you're probably hitting right now.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
Use the discount code AIND-X-50 to get 50% off your tickets.
From endless tool lists to meaningful rankings, Alan Pope, DevRel at Tessl, Simon Maple on AI Native Dev to unpack what defines ‘good’ in an open ecosystem and the tools that have helped him replace manual effort.
Catch their full conversation now. Links in the comments.
Catch their full conversation now. Links in the comments.
November 6, 2025 at 9:21 AM
From endless tool lists to meaningful rankings, Alan Pope, DevRel at Tessl, Simon Maple on AI Native Dev to unpack what defines ‘good’ in an open ecosystem and the tools that have helped him replace manual effort.
Catch their full conversation now. Links in the comments.
Catch their full conversation now. Links in the comments.
At the AI Native DevCon, you'll find workshops with practitioners who are deep in the weeds:
Bring your dev environments, and let's learn together.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
Bring your dev environments, and let's learn together.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
November 6, 2025 at 9:14 AM
At the AI Native DevCon, you'll find workshops with practitioners who are deep in the weeds:
Bring your dev environments, and let's learn together.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
Bring your dev environments, and let's learn together.
📍 DevCon is November 18-19th in NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
𝐓𝐡𝐞 𝐓𝐨𝐨𝐥𝐬-𝐢𝐧-𝐀𝐜𝐭𝐢𝐨𝐧 𝐭𝐫𝐚𝐜𝐤 𝐢𝐬 𝐰𝐡𝐞𝐫𝐞 𝐭𝐨𝐨𝐥𝐬 𝐬𝐭𝐞𝐩 𝐨𝐮𝐭 𝐨𝐟 𝐭𝐡𝐞 𝐬𝐥𝐢𝐝𝐞𝐬 𝐚𝐧𝐝 𝐢𝐧𝐭𝐨 𝐫𝐮𝐧𝐧𝐢𝐧𝐠 𝐜𝐨𝐝𝐞.
Each session is code-on-screen, showing real use cases. If you’re building with or for devs, this is the track you don’t skip
📍 DevCon is November 18-19th, NYC.
👉 Tickets and full lineup: ainativedev.co/z84
Each session is code-on-screen, showing real use cases. If you’re building with or for devs, this is the track you don’t skip
📍 DevCon is November 18-19th, NYC.
👉 Tickets and full lineup: ainativedev.co/z84
November 5, 2025 at 8:48 AM
𝐓𝐡𝐞 𝐓𝐨𝐨𝐥𝐬-𝐢𝐧-𝐀𝐜𝐭𝐢𝐨𝐧 𝐭𝐫𝐚𝐜𝐤 𝐢𝐬 𝐰𝐡𝐞𝐫𝐞 𝐭𝐨𝐨𝐥𝐬 𝐬𝐭𝐞𝐩 𝐨𝐮𝐭 𝐨𝐟 𝐭𝐡𝐞 𝐬𝐥𝐢𝐝𝐞𝐬 𝐚𝐧𝐝 𝐢𝐧𝐭𝐨 𝐫𝐮𝐧𝐧𝐢𝐧𝐠 𝐜𝐨𝐝𝐞.
Each session is code-on-screen, showing real use cases. If you’re building with or for devs, this is the track you don’t skip
📍 DevCon is November 18-19th, NYC.
👉 Tickets and full lineup: ainativedev.co/z84
Each session is code-on-screen, showing real use cases. If you’re building with or for devs, this is the track you don’t skip
📍 DevCon is November 18-19th, NYC.
👉 Tickets and full lineup: ainativedev.co/z84
The NOW track at DevCon - this is the stuff that actually works today. Not demos. Not beta waitlists. Production-ready.
This isn't about what's coming. It's about what's working right now that you can use Monday morning.
📍 DevCon is Nov 18-19th. NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
This isn't about what's coming. It's about what's working right now that you can use Monday morning.
📍 DevCon is Nov 18-19th. NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
November 3, 2025 at 3:02 PM
The NOW track at DevCon - this is the stuff that actually works today. Not demos. Not beta waitlists. Production-ready.
This isn't about what's coming. It's about what's working right now that you can use Monday morning.
📍 DevCon is Nov 18-19th. NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
This isn't about what's coming. It's about what's working right now that you can use Monday morning.
📍 DevCon is Nov 18-19th. NYC.
🎫 Tickets and full lineup: ainativedev.co/z84
𝐓𝐡𝐞 𝐰𝐨𝐫𝐤𝐬𝐡𝐨𝐩 𝐭𝐫𝐚𝐜𝐤 𝐚𝐭 𝐃𝐞𝐯𝐂𝐨𝐧 𝐢𝐬 𝐰𝐡𝐞𝐫𝐞 𝐭𝐡𝐞𝐨𝐫𝐲 𝐦𝐞𝐞𝐭𝐬 𝐤𝐞𝐲𝐛𝐨𝐚𝐫𝐝.
Not demos. but hands-on time with practitioners who are building the AI dev stack right now.
These are working sessions. Bring your dev laptop.
📍 DevCon is November 18-19th in NYC
🎫 Tickets and full lineup: ainativedev.co/z84
#AINDCon
Not demos. but hands-on time with practitioners who are building the AI dev stack right now.
These are working sessions. Bring your dev laptop.
📍 DevCon is November 18-19th in NYC
🎫 Tickets and full lineup: ainativedev.co/z84
#AINDCon
October 31, 2025 at 10:39 AM
𝐓𝐡𝐞 𝐰𝐨𝐫𝐤𝐬𝐡𝐨𝐩 𝐭𝐫𝐚𝐜𝐤 𝐚𝐭 𝐃𝐞𝐯𝐂𝐨𝐧 𝐢𝐬 𝐰𝐡𝐞𝐫𝐞 𝐭𝐡𝐞𝐨𝐫𝐲 𝐦𝐞𝐞𝐭𝐬 𝐤𝐞𝐲𝐛𝐨𝐚𝐫𝐝.
Not demos. but hands-on time with practitioners who are building the AI dev stack right now.
These are working sessions. Bring your dev laptop.
📍 DevCon is November 18-19th in NYC
🎫 Tickets and full lineup: ainativedev.co/z84
#AINDCon
Not demos. but hands-on time with practitioners who are building the AI dev stack right now.
These are working sessions. Bring your dev laptop.
📍 DevCon is November 18-19th in NYC
🎫 Tickets and full lineup: ainativedev.co/z84
#AINDCon
The Edge track at AI Native DevCon is about emerging ideas in AI development that aren't fully polished yet.
These are practitioners working through problems that don't have established solutions yet.
📍 DevCon is November 18-19th in NYC.
👉 Tickets and full lineup: ainativedev.co/z84
These are practitioners working through problems that don't have established solutions yet.
📍 DevCon is November 18-19th in NYC.
👉 Tickets and full lineup: ainativedev.co/z84
October 30, 2025 at 8:54 AM
The Edge track at AI Native DevCon is about emerging ideas in AI development that aren't fully polished yet.
These are practitioners working through problems that don't have established solutions yet.
📍 DevCon is November 18-19th in NYC.
👉 Tickets and full lineup: ainativedev.co/z84
These are practitioners working through problems that don't have established solutions yet.
📍 DevCon is November 18-19th in NYC.
👉 Tickets and full lineup: ainativedev.co/z84
AI models claim to be best at coding, but benchmarks don't tell the full story.
Adam Larson spent 2024-25 building custom tools to test these claims.
Nov 19 at DevCon NYC, Adam will share what actually matters when choosing AI coding tools — beyond the hype.
👉 Register at: ainativedev.co/px8
Adam Larson spent 2024-25 building custom tools to test these claims.
Nov 19 at DevCon NYC, Adam will share what actually matters when choosing AI coding tools — beyond the hype.
👉 Register at: ainativedev.co/px8
October 29, 2025 at 6:16 PM
AI models claim to be best at coding, but benchmarks don't tell the full story.
Adam Larson spent 2024-25 building custom tools to test these claims.
Nov 19 at DevCon NYC, Adam will share what actually matters when choosing AI coding tools — beyond the hype.
👉 Register at: ainativedev.co/px8
Adam Larson spent 2024-25 building custom tools to test these claims.
Nov 19 at DevCon NYC, Adam will share what actually matters when choosing AI coding tools — beyond the hype.
👉 Register at: ainativedev.co/px8
Here are some of our wonderful speakers
Expect deep dives, live demos, and unfiltered conversations on how AI is reshaping the craft of software development, from agentic systems and specs to review workflows and human-in-the-loop design.
Head over to ainativedev.io/devcon to see the full line up!
Expect deep dives, live demos, and unfiltered conversations on how AI is reshaping the craft of software development, from agentic systems and specs to review workflows and human-in-the-loop design.
Head over to ainativedev.io/devcon to see the full line up!
October 28, 2025 at 10:34 AM
Here are some of our wonderful speakers
Expect deep dives, live demos, and unfiltered conversations on how AI is reshaping the craft of software development, from agentic systems and specs to review workflows and human-in-the-loop design.
Head over to ainativedev.io/devcon to see the full line up!
Expect deep dives, live demos, and unfiltered conversations on how AI is reshaping the craft of software development, from agentic systems and specs to review workflows and human-in-the-loop design.
Head over to ainativedev.io/devcon to see the full line up!
Prompts are temporary. Specs are forever.
OpenAI’s Sean Grove calls specs the “new code.” They force you to decide what good output looks like before you run the AI.
Think reusable clauses, version control, IDEs that validate your spec as you write. 🧵
#AINativeDev #AIEngineering
OpenAI’s Sean Grove calls specs the “new code.” They force you to decide what good output looks like before you run the AI.
Think reusable clauses, version control, IDEs that validate your spec as you write. 🧵
#AINativeDev #AIEngineering
July 10, 2025 at 8:40 AM
Prompts are temporary. Specs are forever.
OpenAI’s Sean Grove calls specs the “new code.” They force you to decide what good output looks like before you run the AI.
Think reusable clauses, version control, IDEs that validate your spec as you write. 🧵
#AINativeDev #AIEngineering
OpenAI’s Sean Grove calls specs the “new code.” They force you to decide what good output looks like before you run the AI.
Think reusable clauses, version control, IDEs that validate your spec as you write. 🧵
#AINativeDev #AIEngineering
AI coding tools are powerful, but without structure, they drift.
Claude Taskmaster attempts to fix that, and its growth has been explosive. In just 9 weeks the repo's jumped from 0 → 15,500 stars 🌟
Claude Taskmaster attempts to fix that, and its growth has been explosive. In just 9 weeks the repo's jumped from 0 → 15,500 stars 🌟
July 2, 2025 at 8:37 AM
AI coding tools are powerful, but without structure, they drift.
Claude Taskmaster attempts to fix that, and its growth has been explosive. In just 9 weeks the repo's jumped from 0 → 15,500 stars 🌟
Claude Taskmaster attempts to fix that, and its growth has been explosive. In just 9 weeks the repo's jumped from 0 → 15,500 stars 🌟
Why code alone isn’t enough anymore 🚫💻
Most AI tools today just help us write more code, but the real shift is moving beyond code as the source of truth.
At our first London meetup, we dug into a better approach: spec-driven development.
Most AI tools today just help us write more code, but the real shift is moving beyond code as the source of truth.
At our first London meetup, we dug into a better approach: spec-driven development.
July 1, 2025 at 10:10 AM
Why code alone isn’t enough anymore 🚫💻
Most AI tools today just help us write more code, but the real shift is moving beyond code as the source of truth.
At our first London meetup, we dug into a better approach: spec-driven development.
Most AI tools today just help us write more code, but the real shift is moving beyond code as the source of truth.
At our first London meetup, we dug into a better approach: spec-driven development.
We built Glaser, an AI code reviewer with stand-up comic energy: “48 code blocks? This README is more code than text!”
Structured, snarky AI. Ruby devs, check it out.
🔗 ainativedev.co/lyr
Structured, snarky AI. Ruby devs, check it out.
🔗 ainativedev.co/lyr
June 30, 2025 at 4:04 PM
We built Glaser, an AI code reviewer with stand-up comic energy: “48 code blocks? This README is more code than text!”
Structured, snarky AI. Ruby devs, check it out.
🔗 ainativedev.co/lyr
Structured, snarky AI. Ruby devs, check it out.
🔗 ainativedev.co/lyr
Structure > force.
One structured prompt = same accuracy as eight unstructured ones (via Will Kurt, .txt).
For Claude Code users: clearer specs mean faster output, less context bloat, reusable prompts.
🔗 Read how specs can elevate your coding workflow.
ainativedev.co/bm3
#ClaudeCode #SpecDriven
One structured prompt = same accuracy as eight unstructured ones (via Will Kurt, .txt).
For Claude Code users: clearer specs mean faster output, less context bloat, reusable prompts.
🔗 Read how specs can elevate your coding workflow.
ainativedev.co/bm3
#ClaudeCode #SpecDriven
June 30, 2025 at 7:47 AM
Structure > force.
One structured prompt = same accuracy as eight unstructured ones (via Will Kurt, .txt).
For Claude Code users: clearer specs mean faster output, less context bloat, reusable prompts.
🔗 Read how specs can elevate your coding workflow.
ainativedev.co/bm3
#ClaudeCode #SpecDriven
One structured prompt = same accuracy as eight unstructured ones (via Will Kurt, .txt).
For Claude Code users: clearer specs mean faster output, less context bloat, reusable prompts.
🔗 Read how specs can elevate your coding workflow.
ainativedev.co/bm3
#ClaudeCode #SpecDriven
Agent-based coding ≠ just pair programming anymore.
We’re seeing something bigger:
▪️ Parallel agents handling subtasks
▪️ Isolated envs via git worktrees + containers
▪️ New UIs for merge + review
🔗 Your guide to parallelizing AI coding agents:
ainativedev.co/tg5
#ainativedev #aiagents
We’re seeing something bigger:
▪️ Parallel agents handling subtasks
▪️ Isolated envs via git worktrees + containers
▪️ New UIs for merge + review
🔗 Your guide to parallelizing AI coding agents:
ainativedev.co/tg5
#ainativedev #aiagents
June 20, 2025 at 8:04 AM
Agent-based coding ≠ just pair programming anymore.
We’re seeing something bigger:
▪️ Parallel agents handling subtasks
▪️ Isolated envs via git worktrees + containers
▪️ New UIs for merge + review
🔗 Your guide to parallelizing AI coding agents:
ainativedev.co/tg5
#ainativedev #aiagents
We’re seeing something bigger:
▪️ Parallel agents handling subtasks
▪️ Isolated envs via git worktrees + containers
▪️ New UIs for merge + review
🔗 Your guide to parallelizing AI coding agents:
ainativedev.co/tg5
#ainativedev #aiagents
June 17, 2025 at 7:54 AM
📝 Specs are the new code.
Prompting was just the start. The next level of AI-native development is Specification Driven Development and Sean Grove (OpenAI) nailed why 🧵
Prompting was just the start. The next level of AI-native development is Specification Driven Development and Sean Grove (OpenAI) nailed why 🧵
June 11, 2025 at 2:57 PM
📝 Specs are the new code.
Prompting was just the start. The next level of AI-native development is Specification Driven Development and Sean Grove (OpenAI) nailed why 🧵
Prompting was just the start. The next level of AI-native development is Specification Driven Development and Sean Grove (OpenAI) nailed why 🧵
devs: you might want to try this one 👀
Mistral's new open-source model for coding agents, Devstral, hit 46.8% on SWE-bench verified.
That’s higher than some proprietary giants.
We tested it, check out the write up!
ainativedev.co/fa9
#ainative #opensourceai #devstral
Mistral's new open-source model for coding agents, Devstral, hit 46.8% on SWE-bench verified.
That’s higher than some proprietary giants.
We tested it, check out the write up!
ainativedev.co/fa9
#ainative #opensourceai #devstral
June 10, 2025 at 10:05 AM
devs: you might want to try this one 👀
Mistral's new open-source model for coding agents, Devstral, hit 46.8% on SWE-bench verified.
That’s higher than some proprietary giants.
We tested it, check out the write up!
ainativedev.co/fa9
#ainative #opensourceai #devstral
Mistral's new open-source model for coding agents, Devstral, hit 46.8% on SWE-bench verified.
That’s higher than some proprietary giants.
We tested it, check out the write up!
ainativedev.co/fa9
#ainative #opensourceai #devstral
Connecting your agent to GitHub?
Invariant uncovered a case where a malicious issue can prompt your agent to leak private repo data to a public PR.
No malware involved, just prompt injection.
Invariant uncovered a case where a malicious issue can prompt your agent to leak private repo data to a public PR.
No malware involved, just prompt injection.
June 9, 2025 at 12:22 PM
Connecting your agent to GitHub?
Invariant uncovered a case where a malicious issue can prompt your agent to leak private repo data to a public PR.
No malware involved, just prompt injection.
Invariant uncovered a case where a malicious issue can prompt your agent to leak private repo data to a public PR.
No malware involved, just prompt injection.
GitHub’s Copilot Agent now runs in the cloud; drafting PRs, writing code, and iterating on reviews autonomously.
It respects CI rules and repo protections, and uses server configs (MCP) for tool access.
We got hands-on with the dev experience:
🔗 ainativedev.co/chg
#Build2025 #AIdev #AINativeDev
It respects CI rules and repo protections, and uses server configs (MCP) for tool access.
We got hands-on with the dev experience:
🔗 ainativedev.co/chg
#Build2025 #AIdev #AINativeDev
June 3, 2025 at 9:39 AM
GitHub’s Copilot Agent now runs in the cloud; drafting PRs, writing code, and iterating on reviews autonomously.
It respects CI rules and repo protections, and uses server configs (MCP) for tool access.
We got hands-on with the dev experience:
🔗 ainativedev.co/chg
#Build2025 #AIdev #AINativeDev
It respects CI rules and repo protections, and uses server configs (MCP) for tool access.
We got hands-on with the dev experience:
🔗 ainativedev.co/chg
#Build2025 #AIdev #AINativeDev
Most code models generate one token at a time. Gemini Diffusion doesn’t. It denoises full sequences in parallel, with:
- Major speed gains (~5x over Gemini Flash)
- Built-in error correction
Is it production-ready? Not yet.
Is it interesting? We think so.
➡️ https://www.ainativedev.co/99j
#LLM#AI
- Major speed gains (~5x over Gemini Flash)
- Built-in error correction
Is it production-ready? Not yet.
Is it interesting? We think so.
➡️ https://www.ainativedev.co/99j
#LLM#AI
May 30, 2025 at 2:48 PM
Most code models generate one token at a time. Gemini Diffusion doesn’t. It denoises full sequences in parallel, with:
- Major speed gains (~5x over Gemini Flash)
- Built-in error correction
Is it production-ready? Not yet.
Is it interesting? We think so.
➡️ https://www.ainativedev.co/99j
#LLM#AI
- Major speed gains (~5x over Gemini Flash)
- Built-in error correction
Is it production-ready? Not yet.
Is it interesting? We think so.
➡️ https://www.ainativedev.co/99j
#LLM#AI
Sourcegraph's Amp is now open to everyone! Amp is their new agentic coding tool built for deep prompting, VS Code-native workflows, and real developer feedback loops. Dion Almaer comments on this release: https://www.ainativedev.co/a00
#AINativeDev #AITools #VSCode #AgenticCoding
#AINativeDev #AITools #VSCode #AgenticCoding
May 29, 2025 at 10:40 AM
Sourcegraph's Amp is now open to everyone! Amp is their new agentic coding tool built for deep prompting, VS Code-native workflows, and real developer feedback loops. Dion Almaer comments on this release: https://www.ainativedev.co/a00
#AINativeDev #AITools #VSCode #AgenticCoding
#AINativeDev #AITools #VSCode #AgenticCoding
Gemini Diffusion might be easy to overlook, but it could quietly shift how code gets generated.
Instead of predicting tokens left to right, it refines full sequences in parallel.
- ~1500 tokens/sec.
- Less locked-in errors.
- Better for editing and refactoring.
➡️ https://www.ainativedev.co/99j
Instead of predicting tokens left to right, it refines full sequences in parallel.
- ~1500 tokens/sec.
- Less locked-in errors.
- Better for editing and refactoring.
➡️ https://www.ainativedev.co/99j
May 28, 2025 at 3:45 PM
Gemini Diffusion might be easy to overlook, but it could quietly shift how code gets generated.
Instead of predicting tokens left to right, it refines full sequences in parallel.
- ~1500 tokens/sec.
- Less locked-in errors.
- Better for editing and refactoring.
➡️ https://www.ainativedev.co/99j
Instead of predicting tokens left to right, it refines full sequences in parallel.
- ~1500 tokens/sec.
- Less locked-in errors.
- Better for editing and refactoring.
➡️ https://www.ainativedev.co/99j
12 MCP Servers That Will Make You A More Productive Developer.
MCP servers are like APIs for your agent’s brain. You define the tools. The agent learns to use them. And suddenly, your LLM can run a Docker container or file a GitHub issue.
👉 12 MCP servers to tinker with: www.ainativedev.co/409
MCP servers are like APIs for your agent’s brain. You define the tools. The agent learns to use them. And suddenly, your LLM can run a Docker container or file a GitHub issue.
👉 12 MCP servers to tinker with: www.ainativedev.co/409
May 28, 2025 at 2:24 PM
12 MCP Servers That Will Make You A More Productive Developer.
MCP servers are like APIs for your agent’s brain. You define the tools. The agent learns to use them. And suddenly, your LLM can run a Docker container or file a GitHub issue.
👉 12 MCP servers to tinker with: www.ainativedev.co/409
MCP servers are like APIs for your agent’s brain. You define the tools. The agent learns to use them. And suddenly, your LLM can run a Docker container or file a GitHub issue.
👉 12 MCP servers to tinker with: www.ainativedev.co/409