Scott Dietzen’s Post

Name: AI coding tools are everywhere. Yet they all fall short when scaling to… | Scott Dietzen | 38 comments
Uploaded: 2024-10-24T15:59:01.030Z
Duration: 6 min 30 s
Channel: Scott Dietzen

Scott Dietzen

Reimagining Software Engineering with AI at Augment Code

4mo Edited

AI coding tools are everywhere. Yet they all fall short when scaling to 100s of developers and complex codebases. We’re solving this problem at Augment Code. Today, we’re sharing the first look at our platform: the first developer AI for teams. What makes Augment different? Context. Every Augment feature is context-aware. This means every suggestion, completion, and interaction reflects the components, APIs, and coding patterns in your codebase. For software engineers on professional teams, context really matters. When AI deeply understands your codebase, incredible things happen… 🤝 Onboarding new developers? Done. Augment brings instant answers to every engineer, so new teammates can quickly get up to speed. Use chat to ask questions like, “where do we bootstrap this app?” or “what’s the cadence of releases?” to get your team ramped, fast. 🎛 Constant context switching? Solved. Get instant answers and code completions, right in your IDE. 🏭 Institutional knowledge silos? Eliminated. No more searching through out-of-date or missing documentation, or worse, fielding constant questions. 💾 Legacy code? Handled. Augment deeply understands your entire codebase, regardless of when it was built, who wrote it, or what languages and dependencies it uses. 💻 Augment works as an IDE extension, starting with VSCode or JetBrains. 💪 Teams at Webflow, Kong Inc., Pigment, and more are already building with Augment. The best way to see if Augment works for your team is to try it, for free. Give us your largest repos and most complex code to see what we can do: https://bit.ly/4fjUFxL Thank you to our earliest customers: your feedback, support, and ideas inspire our team daily. 💚 And thank you to our investors Sutter Hill Ventures, Index Ventures, Innovation Endeavors, Lightspeed, and Meritech Capital 🚀

38 Comments

Transcript

Hi, I'm Guy and I want to show you some of the stuff we've been building an augment and how we use it to develop Augment. So I'm gonna play the role of a new developer who just joined Augment, and typically one of the first things I do when I'm looking at a new repository I'm not familiar with is I ask our system to describe it for me. I find that this is a great way to kind of get oriented. And and see at a high level what's going on in the repository. And so we can see that, yes, we have a large mono repo of a company, a project called Augment. That's correct. We have a lot of stuff related to AI code generation and so on and so forth. And we have these different components around Research Services, tools, infrastructure, We use grenades, we use Docker configurations. These are all correct things. This is all correct. And then I can go in ahead and ask it to tell me more about services, for example, services is. Where we store all of our production code. And so in services we can see, OK, what are the high level components. We have an API proxy, we have authentication, we have a content manager. These are all real components in our system. And So what I found is that in this way, within 10 minutes, 5 to 10 minutes, I get a pretty good sense of what the repository is about, what are the high level components? I don't have to go read any redness or rely on any documentation. I just asked the model and it helps me out. Alright, so now let's say that I've received my first ticket to solve, and the ticket has to do with the feature we call code edits. So first I want to show you what code edits are. Codes are where I have, let's say, a piece of code. I can select it and then I can give them all an instruction, let's say rename Foo to bar. And so the model shows me a diffuse of the code before the code after. You can choose to accept it or reject it, and that's our code edit feature. And let's say that now I'm working on the augment repository. I was asked to take these samples which we collect and then export out of a database. And we want to add session ID to all these samples. Session ID is how we track user sessions and so I'm going to start by asking if I want to add a session ID. I'm going to start by asking what data structure holds exported code edits. Because I want to go find that and get at the session ID field to it. And as you can see, because this is a real large, complicated repository, it's not just one data structure that holds them. There are actually several candidates, several different stages of the pipeline that hold them. But in this case, the model was able to figure out that edit data is the first choice. This is actually where we store these code edits. And so I'm going to go into edit datum. And I'm going to start typing and here the completion models kicks in. Our completion model has context awareness just like the chat model. It knows that session ID is something that we commonly use across the system and so easily completes the field for SO. So far so good. I've been able to locate the place where the change first change needs to happen and make the change. But now the other part of the work begins, which is to go and propagate this field throughout the whole system now. Since I made a change to a data class, maybe I can rely on the type checker to go and tell me where all the other places I need to propagate this change on so I can pull up the problems pane in Vscode. And looking at this list that doesn't look too promising because most of these. Errors that are being reported have nothing to do with the change that I made, and at least in my setup, the problems paying for Python is typically very noisy and not something I can rely on to complete my tasks. So instead I'm going to pull up our suggested edits. Pain and. This this feature looks at the changes that I've made so far and then tries to predict across the whole code base what other changes do I need to make in order to complete my task. So it kind of tries to infer what next edits do I need to make and so we can see how I currently I'm in this edit dot PY file. The first change is in this export edit to JSON L file which is where the samples get exported. And so here there is a function that accepts edit data and returns the dictionary. And so I need to add this entry. All right, so that's good. The 2nd place is an interesting 1. So as you can see, this is a SQL query that's inside a string inside a Python file. This is the SQL query that pulls these exported coded samples from the database. And so the model is saying if you added session ID to the data class, probably you also want to add. Who wanna query it from the database and so you need to get this JSON value and then select on it. So I'm going to accept that. Um. Once I change the the query from the database, there's a change to related data structure. This is actually the raw data structure that represents a row in the database. It needs to have session ID as well. So these are all real changes. Once I did that, I can see that there is this constructor here. Once I go to it, I can see that the type checker is complaining that I don't have session ID. The model already knows that we need session ID. Populated with the row field and so on and so forth. This is a unit test code where the model populates the session ID, constructor entry with test session, and here we can go on and on. So at this point we're finding that. Suggested Edits is actually helping us write some of the PRS as we develop suggested edits and other features. So suggests that it's chat with context awareness, completions with context awareness. These are things we use every day. Augment and finding them extremely useful and hoping you will as well.

Scott Dietzen

Reimagining Software Engineering with AI at Augment Code

4mo

Read more in our launch post here: https://bit.ly/4fdJ1UU

12 Reactions

Todd McGee

Executive Engineering Leader Specialized in Innovative Solutions & Software Organizations

4mo

I wonder, Scott, if you aren't forcing the AI to write code like a human and might be forcing some arbitrary constraints on the solution. Where we are now is sort of an uncomfortable in-between place. We want to use AI to write code that conforms to human constraints. But isn't there an eventual goal that the AI takes over the codebase entirely and the constraints and measures we use will change dramatically?

1 Reaction

Marcelo Wesseler

Former VP & GM @ Instacart

4mo

Congratulations! Scott Dietzen and team - Truly the best AI-based coding tool the codem.com team has seen.

6 Reactions

Peter High

President of Metis Strategy, Host of Technovation podcast, columnist at Forbes, author, and keynote speaker

4mo

Congratulations on the remarkable progress you and your team have already made, Scott! Augment Code's story is inspiring. I can't wait to have you tell it to our audience at Metis Strategy Summit on Tuesday in New York!

3 Reactions

Hieu Pham

xAI

4mo

🎉

7 Reactions

Augment Code

4mo

Join our mission to bring joy back to software engineering! We're hiring: https://grnh.se/77b181ac8us

1 Reaction

Edvin T Berhane

Codepoint Fellow @ SHV. Currently at a stealth Incubation.

4mo

As a new engineer on a team, Augment has been incredibly useful in tackling all the questions and problems that come up beyond just generating code. My favorite feature as someone that just joined is how instead of tapping someone on the shoulder, chat is really good at pointing me to where I need to make changes. This happened like 10 times just yesterday, saving me more than an hour of time and many interruptions for my coworkers.

7 Reactions

Dion Almaer

Building AI native dev

4mo

Go Augment Code! It's nice to get this out there, right Scott? :) A loooot of work behind all of this. I love the customer testimonials... it was so fun to hear from them how Augment has helped their development. Very humbling. https://augmentcode.com/customers

10 Reactions

Faruk Sahin

VP Product & Engineering, Pocketlaw

4mo

Congrats Augment Code team! 🎉 We've been using this fantastic product at Pocketlaw for over six months now, and the boost in productivity has been incredible. Keep up the great work! 🚀

5 Reactions

Lior Neumann

Coding + AI

4mo

This is a truly game changing product, made by a team of superstars!

5 Reactions

See more comments

To view or add a comment, sign in

More Relevant Posts

Harjoth Khara

Company Helper
4mo
Report this post
Got access a few weeks ago, and I'm loving it! Fewer prompts (it's deep in your codebase), better results, and lightning-fast—like Groq speed! Plus, it fits right into my VSCode workflow. It's my go-to code assistant now. Engineering Teams, give it a try—Augment Code is secure by default and in SOC-2 mode! https://lnkd.in/g4QHZsH8

Scott Dietzen

Reimagining Software Engineering with AI at Augment Code
4mo Edited

AI coding tools are everywhere. Yet they all fall short when scaling to 100s of developers and complex codebases. We’re solving this problem at Augment Code. Today, we’re sharing the first look at our platform: the first developer AI for teams. What makes Augment different? Context. Every Augment feature is context-aware. This means every suggestion, completion, and interaction reflects the components, APIs, and coding patterns in your codebase. For software engineers on professional teams, context really matters. When AI deeply understands your codebase, incredible things happen… 🤝 Onboarding new developers? Done. Augment brings instant answers to every engineer, so new teammates can quickly get up to speed. Use chat to ask questions like, “where do we bootstrap this app?” or “what’s the cadence of releases?” to get your team ramped, fast. 🎛 Constant context switching? Solved. Get instant answers and code completions, right in your IDE. 🏭 Institutional knowledge silos? Eliminated. No more searching through out-of-date or missing documentation, or worse, fielding constant questions. 💾 Legacy code? Handled. Augment deeply understands your entire codebase, regardless of when it was built, who wrote it, or what languages and dependencies it uses. 💻 Augment works as an IDE extension, starting with VSCode or JetBrains. 💪 Teams at Webflow, Kong Inc., Pigment, and more are already building with Augment. The best way to see if Augment works for your team is to try it, for free. Give us your largest repos and most complex code to see what we can do: https://bit.ly/4fjUFxL Thank you to our earliest customers: your feedback, support, and ideas inspire our team daily. 💚 And thank you to our investors Sutter Hill Ventures, Index Ventures, Innovation Endeavors, Lightspeed, and Meritech Capital 🚀
Like Comment
To view or add a comment, sign in
Lorenzo Thione

OG AI Founder (2003) / Keynote Speaker & Investor / Broadway Producer / 🏳️🌈 Advocate
4mo
Report this post
Ship ship ship! Congrats Gaingels portfolio company Augment Code on launching their code gen platform. Software ate the world, and GenAI is eating software! We are excited to join investors like Sutter Hill Ventures, Index Ventures, Innovation Endeavors, Lightspeed, and Meritech Capital in this investment. But with Github Copilot and Cursor out there, what makes Augment different? Context. Every Augment feature is context-aware – this means every suggestion, completion, and interaction reflects the components, APIs, and coding patterns in your codebase. It's particularly well suited for large enterprise-grade codebases. For software engineers on professional teams, context really matters. https://lnkd.in/gfzF8VTa

Scott Dietzen

Reimagining Software Engineering with AI at Augment Code
4mo Edited

AI coding tools are everywhere. Yet they all fall short when scaling to 100s of developers and complex codebases. We’re solving this problem at Augment Code. Today, we’re sharing the first look at our platform: the first developer AI for teams. What makes Augment different? Context. Every Augment feature is context-aware. This means every suggestion, completion, and interaction reflects the components, APIs, and coding patterns in your codebase. For software engineers on professional teams, context really matters. When AI deeply understands your codebase, incredible things happen… 🤝 Onboarding new developers? Done. Augment brings instant answers to every engineer, so new teammates can quickly get up to speed. Use chat to ask questions like, “where do we bootstrap this app?” or “what’s the cadence of releases?” to get your team ramped, fast. 🎛 Constant context switching? Solved. Get instant answers and code completions, right in your IDE. 🏭 Institutional knowledge silos? Eliminated. No more searching through out-of-date or missing documentation, or worse, fielding constant questions. 💾 Legacy code? Handled. Augment deeply understands your entire codebase, regardless of when it was built, who wrote it, or what languages and dependencies it uses. 💻 Augment works as an IDE extension, starting with VSCode or JetBrains. 💪 Teams at Webflow, Kong Inc., Pigment, and more are already building with Augment. The best way to see if Augment works for your team is to try it, for free. Give us your largest repos and most complex code to see what we can do: https://bit.ly/4fjUFxL Thank you to our earliest customers: your feedback, support, and ideas inspire our team daily. 💚 And thank you to our investors Sutter Hill Ventures, Index Ventures, Innovation Endeavors, Lightspeed, and Meritech Capital 🚀

1 Comment
Like Comment
To view or add a comment, sign in
Jake Kim

Always building | Ex-Engineering @ Apple
4mo Edited
Report this post
Augment Code claims the spot for first developer AI for teams. With 227M raised, it's building a context aware LLM application for engineering that integrates with your IDEs. Applications that layer [context-awareness + vertical-specific integrations] on top of LLMs are going to continue the LLM-wave in the market

Scott Dietzen

Reimagining Software Engineering with AI at Augment Code
4mo Edited

AI coding tools are everywhere. Yet they all fall short when scaling to 100s of developers and complex codebases. We’re solving this problem at Augment Code. Today, we’re sharing the first look at our platform: the first developer AI for teams. What makes Augment different? Context. Every Augment feature is context-aware. This means every suggestion, completion, and interaction reflects the components, APIs, and coding patterns in your codebase. For software engineers on professional teams, context really matters. When AI deeply understands your codebase, incredible things happen… 🤝 Onboarding new developers? Done. Augment brings instant answers to every engineer, so new teammates can quickly get up to speed. Use chat to ask questions like, “where do we bootstrap this app?” or “what’s the cadence of releases?” to get your team ramped, fast. 🎛 Constant context switching? Solved. Get instant answers and code completions, right in your IDE. 🏭 Institutional knowledge silos? Eliminated. No more searching through out-of-date or missing documentation, or worse, fielding constant questions. 💾 Legacy code? Handled. Augment deeply understands your entire codebase, regardless of when it was built, who wrote it, or what languages and dependencies it uses. 💻 Augment works as an IDE extension, starting with VSCode or JetBrains. 💪 Teams at Webflow, Kong Inc., Pigment, and more are already building with Augment. The best way to see if Augment works for your team is to try it, for free. Give us your largest repos and most complex code to see what we can do: https://bit.ly/4fjUFxL Thank you to our earliest customers: your feedback, support, and ideas inspire our team daily. 💚 And thank you to our investors Sutter Hill Ventures, Index Ventures, Innovation Endeavors, Lightspeed, and Meritech Capital 🚀
Like Comment
To view or add a comment, sign in
Rsquared VC

1,084 followers
4mo
Report this post
Augment Code is live! Check out the demo and give it a try at www.augmentcode.com!

Scott Dietzen

Reimagining Software Engineering with AI at Augment Code
4mo Edited

AI coding tools are everywhere. Yet they all fall short when scaling to 100s of developers and complex codebases. We’re solving this problem at Augment Code. Today, we’re sharing the first look at our platform: the first developer AI for teams. What makes Augment different? Context. Every Augment feature is context-aware. This means every suggestion, completion, and interaction reflects the components, APIs, and coding patterns in your codebase. For software engineers on professional teams, context really matters. When AI deeply understands your codebase, incredible things happen… 🤝 Onboarding new developers? Done. Augment brings instant answers to every engineer, so new teammates can quickly get up to speed. Use chat to ask questions like, “where do we bootstrap this app?” or “what’s the cadence of releases?” to get your team ramped, fast. 🎛 Constant context switching? Solved. Get instant answers and code completions, right in your IDE. 🏭 Institutional knowledge silos? Eliminated. No more searching through out-of-date or missing documentation, or worse, fielding constant questions. 💾 Legacy code? Handled. Augment deeply understands your entire codebase, regardless of when it was built, who wrote it, or what languages and dependencies it uses. 💻 Augment works as an IDE extension, starting with VSCode or JetBrains. 💪 Teams at Webflow, Kong Inc., Pigment, and more are already building with Augment. The best way to see if Augment works for your team is to try it, for free. Give us your largest repos and most complex code to see what we can do: https://bit.ly/4fjUFxL Thank you to our earliest customers: your feedback, support, and ideas inspire our team daily. 💚 And thank you to our investors Sutter Hill Ventures, Index Ventures, Innovation Endeavors, Lightspeed, and Meritech Capital 🚀
Like Comment
To view or add a comment, sign in
Siddharth Shah

Founder & CEO at SVECTOR
1mo
Report this post
Introducing SYNC AI CODER: Build, Debug, and Deploy Seamlessly We are thrilled to unveil SYNC AI CODER at SVECTOR, the ultimate AI-powered coding platform that empowers anyone to turn their ideas into reality – no prior coding experience needed. With cutting-edge capabilities powered by the Spec-Coder Model and AGENTIC AI, building apps, websites, and applications has never been this easy or efficient. What Makes SYNC AI CODER Revolutionary? 1. Real-Time GitHub Integration: Push your code directly to GitHub in real time. Save time and streamline your workflow with instant version control. 2. No Coding Experience Required: Now, everyone can bring their ideas to life. Whether you’re a seasoned developer or a beginner, SYNC AI CODER makes it possible to create apps, websites, or any application you can imagine. 3. Dynamic Code Execution and Management: Upload your code folder and run it instantly on the platform. Use Git URLs to fetch and execute entire repositories within SYNC AI CODER. Enjoy full support for integrating and running your projects within the SVECTOR Ecosystem for seamless collaboration and deployment. 4. Effortless Innovation: SYNC AI CODER allows you to build amazing, cool applications that suit your personal or professional needs. Whether you’re developing for business, education, or leisure, the possibilities are endless. Key Features: Own Code Editor for intuitive, hassle-free coding. AI-Powered Terminal for optimized command-line workflows. Automated Bug Fixing to eliminate errors instantly. Real-Time Error Detection and Resolution during project builds. Push-Button GitHub Deployment: Commit your projects effortlessly. High-Level Code Understanding powered by advanced AI models. With SYNC AI CODER, you can unlock the full potential of the SVECTOR Ecosystem to build, manage, and deploy projects faster than ever. Your ideas deserve to shine – let SYNC AI CODER bring them to life. Start building today! #SVECTOR
Like Comment
To view or add a comment, sign in
Benjamin Drury

CTO at DearFlow | Techstars ‘24 | Youngest Students Germany | Passionate Software Engineer
3mo
Report this post
AI makes my coding 50% faster. But how much trust is too much? Using GitHub Copilot has boosted my coding speed by around 40-50%, especially for frontend work. With an existing code base it can use former coded screens and components as context and write the code for similar pages, forms etc. My thoughts for 2025; Software Engineers will shift to platform thinking, focusing on how systems achieve goals, not merely outcomes. (Move closer to PM Role) AI will draft user stories, acceptance criteria, and even basic UI, while we refine strategy and design. (Admittedly; The code will probably suck.) Agile teams will achieve true continuous delivery as AI writes code and drafts PRs at scale. Testing will become more critical, with test architects ensuring AI-generated code works seamlessly. (I’ve been thinking about this a lot. How can I/Should I not care about the code quality in terms of human-readable? If it works and can be changed within seconds or minutes…) Faster workflows, simpler collaboration, more time for innovation. AI won’t replace us—but it will for sure force us to be smarter and in some levels we have to learn to trust it.

6 Comments
Like Comment
To view or add a comment, sign in
Shaswat Raj

CSE'28 BIT Mesra
1mo
Report this post
🚀 Big Launch Alert! Introducing CodeSeek - Your New AI-Powered Coding Companion in VS Code! I’m thrilled to announce the release of CodeSeek, a lightning-fast, context-aware coding assistant built to supercharge your development workflow! 🧠💻 ✨ Why CodeSeek? ✅ DeepSeek AI Integration: Harnesses the power of DeepSeek’s models for smarter, context-aware code completions. ✅ Blazing Speed: Optimized for near-instant suggestions—no more waiting for AI to "think". ✅ Lightweight: Runs smoothly without hogging your RAM (under 50MB!). ✅ Smart Context: Understands your code’s structure, frameworks, and patterns to deliver relevant completions. ✅ Privacy-First: No data collection—your code stays yours. 🚀 Where to Get It: VS Code Marketplace: https://lnkd.in/dVkfU9at Product Hunt Launch: https://lnkd.in/dKZyMeFC GitHub Repo: https://lnkd.in/dXKctgZc 💡 Why I Built This: While exploring AI models for code completion, I discovered DeepSeek—a state-of-the-art model that outperforms many alternatives while costing a fraction of GitHub Copilot, GPT-4, or other closed-source tools. Its speed, accuracy, and affordability blew me away, and I knew it deserved a dedicated, lightweight extension to bring its power to developers everywhere. Let’s democratize access to cutting-edge AI coding tools and build something fast, open, and truly awesome for the world! 🌍 🌟 This is Just the Beginning! CodeSeek is open-source and evolving! Many features are still in the pipeline, and I’d love your help to shape its future: 🔧 Contribute Code: PRs are welcome! 💡 Suggest Features: What’s missing? Debugger integration? Multi-line completions? 🐞 Report Bugs: Let’s squash them together! 👇 How You Can Help: ⭐ Star the GitHub repo to support open-source development! 🚀 Upvote on Product Hunt to help others discover CodeSeek! 💬 Open a GitHub Issue to share ideas or report bugs. 🔄 Tag a dev who loves building tools! > There are many features yet to be added. 🔍 Behind the Scenes: Built with Node.js + VSCode API + DeepSeek API Custom context-trimming algorithm for minimal latency 100% Open Source (MIT Licensed) Your code, your rules 🔒 #AI #Coding #OpenSource #DeveloperTools #VSCode #CommunityDriven #TechInnovation #DeepSeek #DeepSeekR1 #CodeSeek #GitHubCopilot Let’s build the future of coding assistants—together! 🛠️🚀

1 Comment
Like Comment
To view or add a comment, sign in
Rajeev Rajan
3mo
Report this post
AI is bringing joy back to coding. In my recent piece for The New Stack, I shared how AI is helping developers tackle the “laundry and dishes” of software development—tedious tasks like managing tech debt and hunting down documentation—so they can focus on what they love: building. Read more about how AI is transforming the developer experience: https://lnkd.in/gaStFJm3

AI Is Helping Developers Fall in Love With Coding Again

https://thenewstack.io

5 Comments
Like Comment
To view or add a comment, sign in
Jenna Pederson

Developer relations leader | Keynote speaker | Software engineer
7mo
Report this post
Dreading that code review where you find out you didn’t follow your company’s patterns? Instead of digging through code for existing patterns, copying and pasting, looking for APIs you need, or coding it up from scratch, you can use an AI coding companion customized to ✨your✨ unique code base, company’s coding styles, best practices, patterns, and custom libraries and APIs. By doing this, you can: 🔹 keep your code DRY 🔹 improve on-boarding and knowledge transfer 🔹 improve team collaboration Read more about this in my latest blog post ⤵️⤵️⤵️

How an AI coding companion keeps my code DRY

community.aws
Like Comment
To view or add a comment, sign in
Cagdas Cirit

Director of Engineering at Uber
2mo
Report this post
‘The 70% Problem’ in AI-Assisted Coding: A Must-Read by Addy Osmani AI-assisted development tools are transforming the way we code, but are they truly making our software better? Addy Osmani’s latest article, “The 70% Problem: Hard Truths About AI-Assisted Coding,” delves into this pressing question, offering a thought-provoking perspective on how we should rethink our expectations and workflows. This is essential reading for developers and engineering leaders navigating AI’s evolving role in our industry. Grateful to Gergely Orosz for spotlighting this gem and igniting such valuable discussions! Read the full article here: https://lnkd.in/eVY4v-sq

The 70% problem: Hard truths about AI-assisted coding

addyo.substack.com

8 Comments
Like Comment
To view or add a comment, sign in

10,195 followers

213 Posts

View Profile Connect

Scott Dietzen’s Post

Transcript

More Relevant Posts

Explore topics