AI Cheat Codes, WaPo x OpenAI, & The Palantir Divide | Intent, 021

It’s been a few months since we hit your inbox, so if you’re confused — yep, you subscribed at some point!

As a reminder, Intent is all about helping talent in tech become more intentional with their career by staying informed, fluent, and aware of what’s going on in and around the industry. Thanks for sticking with us!

Today's agenda:

• Cluely: The AI app for cheating on… everything?

• WaPo x OpenAI: A new partnership for LLM search content

• Y Combinator's Founder v Palantir: A moral debate

• & three interesting startups

Cluely, the AI tool that wants you to cheat on everything

A 21-year-old Columbia dropout named Chungin “Roy” Lee just raised $5.3 million from VCs for his startup, Cluely. It’s an AI desktop assistant that sees your screen and hears your audio, without the other side of any conversation being aware.

The way the founder pitches it: he wants to help users cheat on exams, job interviews, sales calls, and even dates.

Their polarizing launch video showed someone on a date using the tool to awkwardly lie about his age and art knowledge at a restaurant, drawing apt "Black Mirror" comparisons.

Cluely’s manifesto compares their tool to calculators and spellcheck – innovations once decried as cheating. But let's apply some nuance here. Calculators augment calculation and spell check corrects typos; both operate within established expectations.

Our take:

• Cluely wants to operate in situations where authenticity is the implicit assumption. They want to misrepresent a user’s knowledge or personality.

• Employers, test-givers, and yes, even dates don’t expect you to have covert AI assistance. There’s an unspoken agreement that your knowledge – your authenticity – is what’s on the table.

• Marketing stunts are nothing new. The shock value has definitely worked: going viral on social media, hooking investors, all while creating a swirl of controversy. But stunts can push moral boundaries, especially when they trivialize the notion of cheating.

WaPo Feeds the Beast with an OpenAI Partnership

The Washington Post just announced a new partnership with OpenAI. The deal means ChatGPT users will see summaries, quotes, and links to WaPo articles in response to relevant queries, complete with clear attribution.

For The Washington Post, this is about "meeting audiences where they are," as their Head of Global Partnerships put it.

Translation: it’s about expanding reach beyond their paywall. Every publisher wants to protect their traffic and drive up subscriptions. Plus, they get to reinforce their brand authority within the dominant AI platform. It's a distribution play in an era where platforms eat publishers. Interestingly, WaPo emphasized it remains "LLM-agnostic," despite this deep integration.

For OpenAI, it’s a valuable new data channel (both for enriching responses and for future model training). As LLMs hunger for high-quality, timely information, forging relationships with credible publishers is key — especially in a data-scarce climate (where unique data is increasingly locked behind paywalls).

Paul Graham vs. Palantir: The Moral Debate

Let’s set the scene: a few days ago, Y Combinator founder Paul Graham tweeted that first-rate programmers have “a huge number of other places” to work “rather than at the company building the infrastructure of the police state.”

This was in response to ICE signing a $30 million contract with Palantir to build ‘ImmigrationOS,’ a software platform to expand surveillance and enforcement operations.

Palantir exec Ted Mabrey fired back with an aggressive response, calling the stance a “luxury belief.”

Breakdown:

• Graham sees Palantir as enabling state overreach at a time when the US constitution is clearly being challenged. To him, it’s a moral choice for tech talent.

• Mabrey thinks Graham isn’t taking the struggle of government enforcement seriously and that Graham’s ask for Palantir to commit to not violate the US constitution is in bad faith.

It’s a powerful conversation, made more high stakes than usual thanks to the presence of a fascist leader who can’t seem to go a day without cancelling a student visa or disappearing an immigrant.

There’s a real tension between building powerful tools for the sake of national security and the ethical red lines that those tools might cross. Aside from ICE, companies like Palantir and Anduril do key innovation work for national defense, and there’s no good argument for letting other companies build modern weapons and defense systems while we sit on our hands. 

We think Mabrey and co could commit to the constitution, though.

Three interesting startups to watch (some are hiring!)

• Strawberry — browser that layers AI research, writing, and automation onto any site. Assistants can fill out forms and do quick research in 100s of tabs . Hiring: Founding Engineers (SF, will relo).

• Aqua — extremely fast and extremely accurate speech to text dictation that works in every text field, from Gmail to your terminal. Speech → code at 4× typing speed.

• Supabase — open‑source Postgres backend benefitting from the rise of “vibe coders,” fresh off a $200 M Series D at $2 B valuation. Mission: be the Firebase of the AI era. Hiring: Engineering, Growth & Ops roles (Global Remote).

Think a friend could use a dose of Intent? Forward this along – inbox envy is real.

Sent with Intent,
By Free Agency