Brad Menezes, CEO of enterprise vibe coding startup Superblocks, believes the following crop of billion-dollar startup concepts are hiding in nearly plain sight: the system prompts utilized by present unicorn AI startups.
System prompts are the prolonged prompts — over 5,000-6,000 phrases — that AI startups use to instruct the foundational fashions from corporations like OpenAI or Anthropic on find out how to generate their application-level AI merchandise. They’re, in Menezes view, like a grasp class in immediate engineering.
“Each single firm has a very totally different system immediate for a similar [foundational] mannequin,” he informed TechCrunch. “They’re making an attempt to get the mannequin to do precisely what’s required for a particular area, particular duties.”
System prompts aren’t precisely hidden. Clients can ask many AI instruments to share theirs. However they aren’t at all times publicly out there.
In order a part of his personal startup’s new product announcement of an enterprise coding AI agent named Clark, Superblocks supplied to share a file of 19 system prompts from a few of the hottest AI coding merchandise like Windsurf, Manus, Cursor, Lovable and Bolt.
Menezes’s tweet went viral, considered by nearly 2 million together with huge names within the Valley like Sam Blond, previously of Founders Fund and Brex, and Aaron Levie, a Superblocks investor. Superblocks introduced final week that it raised a $23 million Collection A, bringing its complete to $60 million for its vibe coding instruments geared to non-developers at enterprises.
So we requested Menezes to stroll us by means of find out how to examine different’s system prompts to glean insights.
“I’d say the largest studying for us constructing Clark and studying by means of the system prompts is that the system immediate itself is possibly 20% of the key sauce,” Menezes defined. This immediate provides the LLM the baseline of what to do.
The opposite 80% is “immediate enrichment” he mentioned, which is the infrastructure a startup builds across the calls to the LLM. That half consists of directions it attaches to a consumer’s immediate, and actions taken when returning the response, corresponding to checking for accuracy.
Roles, context and instruments
He mentioned there are three components of system prompts to check: function prompting, contextual prompting, and gear use.
The very first thing to note is that, whereas system prompts are written in pure language, they’re exceptionally particular. “You principally have to talk as when you would to a human co-worker,” Menezes mentioned. “And the directions need to be excellent.”
Function prompting helps the LLMs be constant, giving each goal and character. As an illustration, Devin’s begins with, “You might be Devin, a software program engineer utilizing an actual pc working system. You’re a actual code-wiz: few programmers are as gifted as you at understanding codebases, writing practical and clear code, and iterating in your adjustments till they’re appropriate.”
Contextual prompting provides the fashions the context to contemplate earlier than performing. It ought to present guardrails that may, as an example, scale back prices and guarantee readability on duties.
Cursor’s instructs, “Solely name instruments when wanted, and by no means point out device names to the consumer — simply describe what you’re doing. … don’t present code except requested. … Learn related file content material earlier than enhancing and repair clear errors, however don’t guess or loop fixes greater than 3 times.”
Device use permits agentic duties as a result of it instructs the fashions find out how to transcend simply producing textual content. Replit’s, as an example, is lengthy and describes enhancing and looking out code, putting in languages, organising and querying PostgreSQL databases, executing shell instructions and extra.
Finding out others’ system prompts helped Menezes see what different vibe coders emphasised. Instruments like Loveable, V0, and Bolt “deal with quick iteration,” he mentioned, whereas “Manus, Devin, OpenAI Codex, and Replit” assist customers create full-stack functions however “the output remains to be uncooked code.”
Menezes noticed a possibility to let non-programmers write apps, if his startup may deal with extra, corresponding to safety and entry to enterprise knowledge sources like Salesforce.
Whereas he’s not but operating the multi-billion startup of his desires, Superblock has landed some notable corporations as prospects, it mentioned, together with Instacart and Paypaya World.
Menezes can also be dogfooding the product internally. His software program engineers will not be allowed to jot down inside instruments; they will solely construct the product. So his enterprise people have constructed brokers for all their wants, like one which makes use of CRM knowledge to establish leads, one which tracks assist metrics, one other that steadiness the assignments of the human gross sales engineers.
“That is principally a approach for us to construct the instruments and never purchase the instruments,” he sais.
