Increase Code, an AI coding assistant startup, unveiled its new “Augment Agent” know-how at the moment, designed to deal with the complexity of huge software program engineering tasks reasonably than easy code technology. The corporate claims its strategy represents a major departure from different AI coding instruments by specializing in serving to builders navigate and modify giant, established codebases that span thousands and thousands of strains of code throughout a number of repositories.
The corporate additionally introduced it has achieved the very best rating to this point on SWE-bench verified, an trade benchmark for AI coding capabilities, by combining Anthropic’s Claude Sonnet 3.7 with OpenAI’s O1 reasoning mannequin.
“Most work in the coding AI space, which is clearly a hot sector, has focused on what people call ‘zero to one’ or ‘vibe coding’ – starting with nothing and producing a piece of software by the end of the session,” stated Scott Dietzen, CEO of Increase Code, in an unique interview with VentureBeat. “What we targeted instead is the software engineering discipline of maintaining big, complex systems — databases, networking stacks, storage — codebases that have evolved over many years with hundreds of developers working on them collaboratively.”
Based in 2022, Increase Code has raised $270 million in complete funding, together with a $227 million Sequence B spherical introduced in April 2024 at a post-money valuation of $977 million. The corporate’s buyers embrace Sutter Hill Ventures, Index Ventures, Innovation Endeavors (led by former Google CEO Eric Schmidt), Lightspeed Enterprise Companions, and Meritech Capital.
How Increase’s context engine tackles multi-million line codebases
What units Increase Agent aside, in accordance with the corporate, is its potential to grasp context throughout large codebases. The agent boasts a 200,000 token context window, considerably bigger than most opponents.
“The challenge for any AI system, including Augment, is that when you’re working with large systems containing tens of millions of lines of code – which is typical for meaningful software applications – you simply can’t pass all that as context to today’s large language models,” defined Dietzen. “We’ve trained our AI models to perform sophisticated real-time sampling, identifying precisely the right subset of the codebase that allows the agent to do its job effectively.”
This strategy contrasts with opponents that both don’t deal with giant codebases or require builders to manually assemble the related context themselves.
One other differentiator is Increase’s real-time synchronization of code modifications throughout groups. “Most of our competitors work with stale versions of the codebase,” stated Dietzen. “If you and I are collaborating in the same code branch and I make a change, you’d naturally want your AI to be aware of that change, just as you would be. That’s why we’ve implemented real-time synchronization of everyone’s view of the code.”
The corporate reviews its strategy has led to a 70% win charge in opposition to GitHub Copilot when competing for enterprise enterprise.
Why ‘memories’ characteristic helps AI match your private coding model
Increase Agent features a “Memories” characteristic that learns from developer interactions to higher align with particular person coding types and preferences over time.
“Part of what we wanted to be able to deliver with our agents is autonomy in the sense that you can give them tasks, but you can also intervene,” Dietzen stated. “Memories are a tool for the model to generalize your intent, to capture that when I’m in this situation, I want you to take this path rather than the path that you took.”
Opposite to the notion that coding is only mathematical logic with out stylistic components, Dietzen emphasised that many builders care deeply concerning the aesthetic and structural features of their code.
“There is definitely a mathematical aspect to code, but there’s also an art to coding as well,” he famous. “Many of our developers want to stay in the code. Some use our agents to write all of the code, but there’s a whole group of engineers that care about what the ultimate code looks like and have strong opinions about that.”
Enterprise adoption of AI coding instruments has been slowed by issues about mental property safety and safety. Increase has centered on addressing these points with a sturdy safety structure and enterprise-grade integrations.
“Agents need to be trusted. If you’re going to give them this autonomy, you want to make sure that they’re not going to do any harm,” stated Dietzen. “We were the first to offer the various levels of SOC compliance and all of the associated penetration testing to harden our solution.”
The corporate has additionally established integration with developer instruments like GitHub, Linear, Jira, Notion, Google Search, and Slack. In contrast to some opponents that implement these integrations on the consumer facet, Increase handles these connections within the cloud, making them “easily shareable and consistent across a larger team,” in accordance with Dietzen.
Increase Agent is mostly accessible for VS Code customers beginning at the moment, with early preview entry for JetBrains customers. The corporate maintains full compatibility with Microsoft’s ecosystem, not like competitor Cursor, which forked VS Code.
“At some level, customers that choose Cursor are opting out of the Microsoft ecosystem. They’re not allowed to use all of the standard VS Code plug-ins that Microsoft provides for access to their environment, whereas we’ve preserved 100% compatibility with VS Code and the Microsoft ecosystem,” Dietzen defined.
The evolving partnership between human engineers and AI assistants
Regardless of the advances in AI coding help, Dietzen believes human software program engineers will stay important for the foreseeable future.
“The arguments around whether software engineering is a good discipline for people going forward are very much off the mark today,” he stated. “The discipline of software engineering is very, very different in terms of crafting and evolving these large code bases, and human insight is going to be needed for years to come.”
Nevertheless, he envisions a future the place AI can tackle extra proactive roles in software program improvement: “The real excitement around where we can ultimately get to with AI is AI just going in and assessing quality of a code base and suggesting how that code base could evolve to make it more secure, more capable, easier to use, more robust.”
Increase Code at present counts Webflow, Pigment, Lemonade, ddn, Datastax, and GoFundMe amongst its prospects. Pricing begins at $30 per person monthly for skilled use and $60 per energetic person monthly for enterprise use, with a free tier additionally accessible.
The launch of Increase Agent comes at a time when enterprises are more and more trying to leverage AI to enhance developer productiveness whereas sustaining software program high quality. As organizations wrestle with sustaining and evolving advanced software program programs, instruments that perceive the nuances of huge codebases could discover a receptive viewers amongst technical leaders looking for to maximise the affect of their engineering groups.
“I’m hopeful that we can actually ramp investment in human software engineering along with machines,” Dietzen concluded, “because we’re going to be able to make much better software, a lot more of it, but much higher quality. And I think we can unleash a tremendous amount of value for humankind by removing all these barriers to great software.”
Each day insights on enterprise use instances with VB Each day
If you wish to impress your boss, VB Each day has you coated. We provide the inside scoop on what firms are doing with generative AI, from regulatory shifts to sensible deployments, so you may share insights for max ROI.
An error occured.
