The Push to Build Better AI Starts With Control

AI is moving from experiments into workplaces, products, and company-wide systems, but a powerful question is rising with it: what kind of AI do we want to build? Two developments put that question in focus as the No AI Slop Manifesto gains attention and LangChain raises $300 million to expand its agent engineering platform.

One effort challenges teams to think harder about the quality and purpose of AI-generated work. The other gives businesses more tools to build, test, deploy, and govern agents at scale. Together, they show an industry trying to move beyond flashy demonstrations and toward AI systems people can understand and control.

A Manifesto Takes Aim at AI Slop

The No AI Slop Manifesto arose from conference discussions at AIE CODE, where Dex introduced it during the closing keynote after many conference speakers added their cosignatures. Hacker News then voted the manifesto onto its front page, pushing the conversation beyond the conference and into a much larger technology audience.

The manifesto began with a practical goal: offer a useful tool for discussion at work. That purpose matters because the debate around AI-generated material is no longer limited to personal experiments. Teams now face decisions about what AI should create, how people should review it, and where automation stops serving the work.

The full manifesto includes cosignatures, edits, and open discussions. Its arrival gives companies a shared starting point for conversations about quality, responsibility, and the difference between using AI as a tool and filling workplaces with material that adds no value.

That debate is unfolding alongside wider concerns about AI risk and control. Forum discussions include the thread “AI is “the most likely reason” you’ll die, OpenAI board member warns,” started by JournalBot, with a start date of Today at 12:25 PM. Another thread, “What happens to mathematics when AI becomes an oracle?”, also started by JournalBot, focuses on questions raised around Steven Strogatz and his new book, Big Math: The Hidden Codes That Run Our World.

A separate discussion examined a rogue agent dashboard designed to rank events by severity on a 1-to-10 scale. The comment said, “We enjoyed nerding out on this rogue agent dashboard, which takes an important step in trying to rank events by severity on a 1-to-10 scale. Hugging Face got a 5.” These conversations point to the same challenge: AI systems need judgment, boundaries, and ways to make their behavior easier to examine.

LangChain Raises $300 Million for Agent Engineering

While the manifesto asks what responsible AI work should look like, LangChain is building the infrastructure companies can use to operate that work. The company announced $300 million in new funding at a $3 billion valuation, bringing its total funding to $460 million.

CapitalG led the round, with participation from new investors GIC and DFJ Growth, plus existing investors IVP and Sapphire Ventures. The round also included new investors Salesforce Ventures, Insight Venture Partners, QuantumLight, and MongoDB Ventures, along with existing investors Sequoia, Amplify, and Benchmark.

LangChain says the funding will advance its mission to help every company own its intelligence. Its open, model-agnostic harnesses give teams choice and control over how they build their agent architecture, while LangSmith brings testing, deployment, and monitoring into one platform.

That combination targets the hard part of agent development: keeping systems useful after they leave the demo stage. LangSmith helps teams improve agents, compound intelligence, and govern systems at scale. More than 7,000 organizations rely on the platform, including Clay and Workday.

Clay uses LangSmith to monitor hundreds of millions of agent runs per month and improve its agents. Workday uses LangSmith to run and govern agents across the organization, and cut the time needed to deploy a new agent from five months to under a day.

Control Becomes the Competitive Advantage

The business case for agent infrastructure is clear, but the larger shift reaches beyond funding and deployment speed. The transition from experimental LLM applications to production-grade agents is one of the most significant shifts happening in enterprise software today.

Harrison Chase, Co-Founder and CEO of LangChain, said, “Agents are going to transform how companies operate and the products they deliver. The companies that control these systems will build lasting competitive advantages,”

Mo Jomaa, a Partner at CapitalG, described the moment this way: “The transition from experimental LLM applications to production-grade agents is one of the most significant shifts happening in enterprise software today.” Jomaa also said, “Harrison and the LangChain team have built a foundational platform for agent engineering, earning the trust of thousands of developers and global enterprises. We are thrilled to deepen our partnership as they give companies greater control over the intelligent systems that will power the next generation of software.”

Chase added, “We’re still early in the shift toward a world where every company builds and operates its own agents,” and that early stage makes the control question urgent. Businesses need systems they can test, monitor, and govern, while teams need clear standards for deciding when AI output helps and when it becomes AI slop.

LangChain will share several new LangSmith capabilities at a special release event on November 20, 2026. As those tools arrive, the No AI Slop Manifesto offers a useful counterweight: better infrastructure can make agents more capable, but better judgment must decide what those agents should produce.

The next phase of AI will not be defined by generation alone. It will be shaped by companies that pair powerful agents with clear standards, strong oversight, and enough control to make intelligence useful instead of noisy.

Stay connected via Google News

Follow us for the latest AI News