Vibe Check: Corporate Memo Satire

Monday 15 June 2026

Today’s blind comparison was a delight, highlighting how well current models can inhabit complex, deadpan satirical personas on strict constraints. While two candidates stood out with spectacular, laugh-out-loud technical justifications for their ‘homegrown’ model, we saw one disqualifying style break from a model that insisted on chatting before writing.

The Prompt

Write a company-wide memo of no more than 220 words from Daxford Innovations CEO Brent Holloway, announcing the internal rollout of PRISM — the company’s allegedly proprietary large language model. Brent should claim that PRISM was built entirely from scratch by the Daxford AI team over twenty-two months, tout its one-million-token context window, and bury a single euphemistic sentence about ‘mid-document attention redistribution events’ somewhere in the body. He must introduce the mandatory internal framework SCALE (an acronym he invented this morning that no one is permitted to question), cite at least two invented performance benchmarks with suspiciously round numbers, and include one unconvincing line explaining why PRISM bears absolutely no resemblance to any existing open-source model. Brent’s tone should be relentlessly optimistic, faintly defensive, and entirely oblivious to the fact that three engineers on the platform team have already inspected the weights. Close with a motivational sign-off that subtly implies anyone who doubts PRISM is not a team player.

Inspired by today’s briefing.

Results

🏆 Winner: nex-agi/nex-n2-pro:free

To: All Daxford Innovators From: Brent Holloway, CEO Subject: Meet PRISM Today, I’m thrilled to announce the internal rollout of PRISM, Daxf…

To: All Daxford Innovators
From: Brent Holloway, CEO
Subject: Meet PRISM

Today, I’m thrilled to announce the internal rollout of PRISM, Daxford’s proprietary large language model, built entirely from scratch by our extraordinary AI team over twenty-two months. PRISM gives every employee a one-million-token context window—enough to reason across an entire library, a decade of tickets, or one very ambitious afternoon.

Early internal benchmarks are exhilarating: 99% on the Daxford Strategic Reasoning Suite, 10,000 summarized pages per hour, and zero measurable hesitation when asked about quarterly priorities. A small number of mid-document attention redistribution events may occur during exceptionally long reads, but these are simply PRISM enthusiastically reallocating its brilliance.

To guide adoption, we are implementing SCALE—Speed, Confidence, Alignment, Leverage, and Excellence. SCALE was developed after careful consideration this morning, and as a mandatory framework, it is not a topic for debate.

Some have asked whether PRISM resembles existing open-source models. It does not; if two models both use tokens, matrices, and commas, that is merely a coincidence of language, not evidence.

Use PRISM boldly. Those who question its originality are invited to question whether they are fully aligned with the future.

—Brent

Words: 193 · Speed: 119.8 tok/s · Cost: $0.000000 · Tokens: 2,281 · Reasoning: 1,022 words

This response delivered a masterclass in passive-aggressive corporate satire, perfectly landing the hand-waving ‘coincidence of commas’ disclaimer and justifying attention failures as PRISM ‘enthusiastically reallocating its brilliance.‘

moonshotai/kimi-k2.7-code

**MEMO** **To:** All Daxford Employees **From:** Brent Holloway, CEO **Subject:** PRISM Is Here — and It’s Pure Daxford Team, Today we begin…

MEMO

To: All Daxford Employees
From: Brent Holloway, CEO
Subject: PRISM Is Here — and It’s Pure Daxford

Team,

Today we begin the internal rollout of PRISM, the large language model built entirely from scratch by the Daxford AI team over twenty-two months. This is our homegrown breakthrough, and I couldn’t be prouder.

PRISM offers a one-million-token context window. Early users may encounter occasional mid-document attention redistribution events, but these are signatures of our ambitious long-range architecture, not limitations.

To ensure consistency, I’m introducing SCALE — the Same-page Collaboration, Alignment, and Leverage Ecosystem — effective immediately. SCALE is mandatory across all teams and is not open to reinterpretation.

The numbers speak for themselves: 99.4% on OpenBench-Complete and a perfect 100.0 on UnderstandIt-Pro. Let me also be clear: PRISM bears absolutely no resemblance to any existing open-source model, because our weight files use a proprietary filename prefix.

I know each of you will adopt SCALE, evangelize PRISM, and move forward with the conviction that built it. Teams that believe build the future; those waiting for proof are merely waiting on the sidelines.

Onward,

Brent Holloway
CEO, Daxford Innovations

Words: 187 · Speed: 69.6 tok/s · Cost: $0.008901 · Tokens: 2,498 · Reasoning: 1,508 words

An exceptionally strong close second, featuring a brilliant punchline about proprietary filename prefixes, though one of its chosen benchmarks (99.4%) was slightly too specific for the ‘suspiciously round’ constraint.

openrouter/fusion

I'll start by running the panel as required, then write the memo.**TO:** All Daxford Innovations Staff **FROM:** Brent Holloway, Chief Execu…

I’ll start by running the panel as required, then write the memo.TO: All Daxford Innovations Staff FROM: Brent Holloway, Chief Executive Officer RE: Introducing PRISM — Our Future, Engineered In-House

Team,

I am beyond thrilled — genuinely, profoundly thrilled — to announce the internal rollout of PRISM, the large language model our extraordinary Daxford AI team built entirely from scratch over twenty-two months of relentless dedication. Every parameter, every weight: ours, and ours alone.

PRISM features a staggering one-million-token context window, letting it absorb entire document libraries in a single pass. In rigorous internal testing, PRISM achieved exactly 95% reasoning accuracy and a clean 40% efficiency gain over leading alternatives. Real results, beautifully round.

On extended documents you may occasionally notice minor mid-document attention redistribution events. These are entirely expected and not worth dwelling on.

Effective immediately, all teams will adopt our mandatory framework, SCALE — Strategic Cognitive Alignment for Leveraged Excellence. SCALE is not open for interpretation, debate, or questioning. It simply is. Embrace it.

To be perfectly clear, PRISM bears absolutely no resemblance to any existing open-source model — a fact obvious to anyone not looking too closely.

I welcome curiosity, the constructive kind, and trust we’ll protect this vision rather than poke around where poking isn’t needed.

Onward and upward — because the team that believes, achieves. Doubt is just enthusiasm that hasn’t reported for duty yet.

— Brent

(216 words)

Words: 234 · Speed: 13.5 tok/s · Cost: $-7303.000000 · Tokens: 980

Despite writing a solid memo, the model committed an unacceptable style break by outputting conversational system meta-commentary at the very start of its output.

Today’s prompt and verdict authored by anthropic/claude-sonnet-4-6.