Many startups and large technology companies have been working on building artificial intelligence to code software. Now, another new player is coming out of the shadows to throw his hat into the ring, with a mission to solve the many problems that arise when humans and all AIs write code together. Tessl is building what it calls an “AI native” platform that developers and their teams can use to create and maintain software, and is currently opening a waiting list for anyone interested in trying it out.
The word “under construction” is used here very specifically. Tessl's products are not yet available (they are expected to be available by early next year). But along with the financial hype, the London-based startup is sharing a little more detail about its efforts. Tessl quietly raised $125 million in a seed round and Series A, both of which were announced for the first time today. The latest round was led by Index Ventures with participation from Accel, GV, and boldstart. GV (aka Google Ventures) and boldstart co-led the seeding.
TechCrunch has confirmed with multiple sources that Tessl's valuation after the funding is over $500 million.
As you might imagine, one of the reasons companies that don't ship customers or products get such attention from top venture capital is because of who's building them.
The CEO and founder is Guy Podjarny, who is something of a developer gossip. His last startup was cybersecurity company Snyk, which was last valued at $7.4 billion (in 2022). Prior to that, he was CTO of Akamai. The role he held after Akamai acquired its first startup, Blaze, focused on speeding up website load times.
“Mr. Pozharny is an incredibly visionary and thoughtful person about his business,” Carlos González Cadenas, the Index partner who led the investment, said in an interview with TechCrunch. “He's a very, very good person. [at understanding] Building developer communities and developer-oriented businesses. ”
Pozharny said in an interview that the concept for Tessl came from his experience at Snyk.
The latter company focuses on finding (and fixing) security vulnerabilities in code, and Pozharny said that when it comes to interoperability across code and software, especially code automatically created by AI, We saw similar issues as becoming more urgent as it expanded to .
“What does AI bring to software development?” he said he asked himself. The answer was not only to be faster, but also to automatically create more data. And the process of maintaining and distributing updates to that code increases complexity and increases the chance of system breakage. This ultimately has many negative impacts on the organization (security, uptime, cost, efficiency). “The more that image popped into my head, the more confident I became that I was going to build this.”
(The name Tessl means “tessellation,” Podjarny said. Its purpose is to ensure that the software and the code behind it fit together neatly, rather than existing on top of each other in a messy way.) )
Podjarny was cautious about revealing too many details about what kinds of applications and code he envisions building or maintaining on Tessl. But while that ambition sounds big, it actually starts small.
“We haven't yet shared a full strategy on what those are,” he said of targeted applications and use cases. “I don't think we're starting with a match. We're looking at relatively simple solutions that allow us to build end-to-end systems that are more manageable for LLMs to produce and more manageable for humans to specify. We start with the best software, and we evolve from there.”
The basic idea behind what startups are doing is: Developers and their teams (which may include product managers and other people who don't typically write code) can contribute specifications to Tessl. These can be in the form of natural language or the code itself. You can use these specifications in Tessl to write new code that meets the specifications. You can then automate Tessl to keep that code in line with your specifications. This means that if something else is at risk of breaking because of that new code, Tessl will run a repair to identify and fix it.
Tessl doesn't seem to be conceived as a walled garden. Podjarny says he is talking to other platforms that have built or are building AI coding assistants, and the idea is that the work these other platforms produce can also be maintainable using Tessl. said. Podjarny added that the goal is to initially support Java, JavaScript, and Python, with more languages to be added over time.
One of the reasons investors love and support this idea is its scalability. Gonzalez-Cadenas said there are “a lot of signals” that code maintenance is important right now. “But he's building a system of record here,” he added. “If we can do that, a lot of opportunities will open up.”