Mert

My top 3 takeaways from the IO 26’ developer keynote, in no particular order:

1. Gemini 3.5 Flash GA is a breakthrough for developers worldwide. In plain English, this model offers flagship-tier coding brains that run at lightning speeds for a fraction of the cost of frontier models. This is a massive win for international devs, as the ultra-low cost levels the playing field against steep USD exchange rates, and the raw speed easily offsets regional network latency. I tested it out myself by vibe coding a vaporwave-themed pet store app with Antigravity. I was impressed with how snappy the code generation felt.

1. Google AI Studio new features can make anyone a builder. One-click deployment to Firebase and Cloud Run makes it easier than ever to go from demo to serving real users. Custom asset generation with Nano Banana makes your vibe coded apps look sharp. And when you’re ready to add next-level complexity, now it’s easier than ever to export your project to Antigravity.

2. Antigravity 2.0 unlocks full-stack agentic development. You can now use always-on agents that build for you while you sleep, while the new Antigravity SDK lets you richly customize your agents and deploy them to private infrastructure.

For more details, check out the official announcement blog and stay tuned for upcoming Dev.To challenges where you can build with our new models. What are you most excited about?

March 2026. Atlassian announces it’s laying off 1,600 people — 10% of its workforce. The announcement drops on a Wednesday. The consultation period ends March 19. April 2 is the last day for hundreds of engineers, designers, and managers who built the tools millions of people use worldwide.

That same quarter, Atlassian reports $1.79 billion in revenue. An all-time record.

Among those let go is a Greek engineer who spent nearly 8 years at the company. His name is Vasilios Syrakis. Senior Systems Engineer. If you use Jira, Confluence, or Bitbucket, there’s a very high probability your traffic passed through infrastructure he designed.

Syrakis doesn’t post a rant. He doesn’t leak proprietary code. He doesn’t expose data.

He posts a 38-minute YouTube video.

Within 8 days, it has 1.1 million views.

Who Is Vasilios Syrakis and What Did He Build?

Syrakis joined Atlassian when the company was a fraction of its current size. Over 8 years, he designed and built systems that became the company’s central nervous system:

Open Service Broker — a web application that let any internal development team provision their own load balancers without asking the infrastructure team for permission. FastAPI, Amazon SQS for task queues, DynamoDB for status tracking. Self-service before self-service was ubiquitous.

Sovereign — the Envoy control plane that replaced expensive enterprise load balancers with an open-source proxy. Syrakis built the management server that dynamically configured Envoy across the entire fleet. He open-sourced it after leaving. 2,000 proxy servers running across 13 AWS regions under this system.

Authentication sidecar in Rust — he wrote it himself. It ran alongside every backend so individual teams didn’t have to build their own security layer. Authentication, access logging, rate limiting — shared services wrapping Jira, Confluence, and Bitbucket.

When Atlassian moved its flagship products behind this infrastructure, Syrakis and his team had already built what was essentially an internal self-service platform that no other person in the company understood end-to-end.

What the Video Actually Showed

Let’s be precise here: Syrakis didn’t leak secrets. He didn’t publish proprietary source code. He didn’t expose customer data.

What he showed was architectures, design patterns, open-source tools, technical decisions, and lessons learned. He explained how real infrastructure is built at scale using accessible tools. The video is deeply technical, professional, and even nostalgic.

The problem isn’t what he showed. The problem is that Atlassian had no idea how much it depended on him.

The video isn’t an act of revenge. It’s a career retrospective. Syrakis also talked about personal growth — how he learned to handle conflict, to mentor, and the challenge of keeping complex systems maintainable as teams and codebases evolve.

But when your career retrospective includes the blueprints of the infrastructure powering products used by millions, you’re doing more than telling your story. You’re documenting — publicly — what your employer never bothered to document internally.

How Atlassian Handled the Layoff

On March 11, 2026, CEO Mike Cannon-Brookes sent an internal memo. The key lines:

“It’s the right decision for Atlassian. But that doesn’t mean it’s easy.”

“Our approach isn’t ‘AI replaces people.’ But it would be disingenuous to pretend AI doesn’t change the mix of skills we need.”

Simultaneously, he announced that CTO Rajeev Rajan was stepping down, to be replaced by two “next-generation AI talents.” The union Professionals Australia stated that employees “had been made redundant without being consulted or given any sign a restructure would affect their jobs.”

The severance package was technically generous: 16 weeks minimum pay, extended healthcare, early pro rata bonus, $1,000 tech payment upon returning the corporate laptop. Slack left open an extra 6 hours for farewells.

But the treatment reveals the mentality: everyone is interchangeable. Same package for 1,600 people. Same process. Same disconnect between what each person knew and what the company retained.

The 5 Leadership Failures That Made the Video Possible

It wasn’t an accident that Syrakis could record 38 minutes of critical infrastructure and publish it without legal consequences. It was the result of leadership failures that repeat across companies of every size:

1. They laid people off without assessing knowledge risk. In a cut of 1,600 people, nobody asked case by case: “what does this person know that nobody else knows?” Syrakis wasn’t a generic employee. He owned the knowledge of systems processing traffic for millions of users. Nobody treated him that way.

2. They treated people like spreadsheet rows. Same package. Same process. Same message. A company that treats the person who built the central nervous system the same as someone who joined 8 months ago is a company that doesn’t know what it has.

3. No handover or transition. From announcement to departure: 3 weeks. No knowledge transfer plan. No guided documentation. No handoff sessions. The knowledge walked out with the person.

4. The timing was perceived as cynical. Record revenue. Profits rising. Investors applauding the cut — the stock went up 4%. Laying off 10% of your people while billing more than ever communicates one thing: people are a cost to trim in order to invest elsewhere. And when you communicate that, don’t expect post-departure loyalty.

5. They replaced the CTO simultaneously. The top technology leader leaves at the same time as 900 R&D engineers. That’s not strategy. It’s a signal that not even leadership knows who built what.

What a Smart Leader Does Differently

Not every company can avoid layoffs. But every company can avoid the scenario where an employee with 8 years of deep knowledge leaves and publishes the entire architecture on YouTube. Here are 6 practices that work:

1. Knowledge dependency map. Before touching the team, you know who knows what. It’s not an HR document. It’s a living map: which system depends on which person? Who’s the only one that understands X? If the answer to that last question isn’t “nobody,” you have a risk.

2. Surgical layoffs with handover. If you have to let go of someone who owns unique knowledge, the severance package should include “X weeks of paid handover.” The company buys the knowledge before it leaves. Cheaper than rebuilding from scratch — or watching it on YouTube.

3. The technical exit interview rule. Before someone leaves — whether laid off or resigning — one technical session where they document what only they know. It’s not optional. It’s part of offboarding. If nothing exists that only you know, great. If something does, it gets documented.

4. Culture of living documentation. Not 300-page manuals nobody reads. Runbooks. Diagrams. Architecture decisions with context. Maintained by the team as part of the job, not as a special project abandoned by sprint 3.

5. Ownership rotation. Nobody should be the sole owner of a critical system for more than 2 years without a second-in-command. If one person leaves and the system is orphaned, the error isn’t the person’s — it’s the leader who allowed a single point of failure.

6. Treat departing employees as allies, not enemies. Syrakis didn’t set out to destroy Atlassian. He set out to tell his story. He could have been a company ambassador. He could have been an external consultant. He could have trained his replacement. They turned him into an accidental technical whistleblower by treating him like a number.

How Guayoyo Tech Can Help

At Guayoyo Tech, we don’t give leadership talks. We build the systems that make institutional knowledge independent of three people’s memory.

Technical knowledge audit: We map which systems depend on which people on your team. We identify single points of human failure before a departure exposes them.

Living documentation implementation: We set up runbook systems, architecture diagrams as code, and wikis your team actually maintains — because we automate them into the pipeline.

Technical offboarding processes: We design handover flows that protect knowledge: guided transfer sessions, architecture decision documentation, and ownership transfer before departure.

Ownership rotation in engineering teams: We implement practices so no critical system has a single owner. Rotating pair programming, cross-reviews, designated seconds-in-command.

You don’t need to be Atlassian for this to happen to you. You just need one person with 5 years of unique knowledge, one leader who doesn’t know what they have, and one layoff round that treats everyone the same.

The difference between an ambassador and an exposer is how you treat them when they leave.

At Guayoyo Tech, we help companies protect their institutional knowledge with systems, not hope.

Book a free discovery call. We map your knowledge risks and tell you exactly what to protect, no commitment.

Marzo de 2026. Atlassian anuncia que va a despedir a 1,600 personas — el 10% de su fuerza laboral. El anuncio llega en miércoles. La consulta termina el 19 de marzo. El 2 de abril es el último día para cientos de ingenieros, diseñadores y gerentes que construyeron las herramientas que usan millones de personas en todo el mundo.

Ese mismo trimestre, Atlassian reporta $1.79 billones de dólares en revenue. Récord histórico.

Entre los despedidos hay un ingeniero griego que lleva casi 8 años en la empresa. Se llama Vasilios Syrakis. Es Senior Systems Engineer. Si usas Jira, Confluence o Bitbucket, hay una probabilidad altísima de que tu tráfico haya pasado por infraestructura que él diseñó.

Syrakis no publica un rant. No filtra código propietario. No filtra datos.

Publica un video de 38 minutos en YouTube.

A los 8 días, tiene 1.1 millones de views.

¿Quién es Vasilios Syrakis y qué construyó?

Syrakis entró a Atlassian cuando la empresa tenía una fracción del tamaño actual. Durante 8 años, diseñó y construyó sistemas que se volvieron el sistema nervioso central de la compañía:

Open Service Broker — una aplicación web que permitía a cualquier equipo interno de desarrollo crear sus propios balanceadores de carga sin pedir permiso al equipo de infraestructura. FastAPI, Amazon SQS para colas de tareas, DynamoDB para tracking de estado. Self-service antes de que self-service fuera ubicuo.

Sovereign — el Envoy control plane que reemplazó balanceadores enterprise carísimos con un proxy open-source. Syrakis construyó el servidor de gestión que configuraba Envoy dinámicamente en toda la flota. Lo open-sourceó después de irse. 2,000 servidores proxy corriendo en 13 regiones de AWS bajo este sistema.

Sidecar de autenticación en Rust — lo escribió él mismo. Corría junto a cada backend sin que cada equipo tuviera que implementar su propia capa de seguridad. Autenticación, logging de accesos, rate limiting — servicios compartidos que envolvían Jira, Confluence y Bitbucket.

Cuando Atlassian movió sus productos principales detrás de esta infraestructura, Syrakis y su equipo ya habían construido lo que esencialmente era una plataforma interna de autoservicio que ninguna otra persona en la empresa entendía de punta a punta.

Lo que realmente mostró el video

Aquí hay que ser precisos: Syrakis no filtró secretos. No publicó código fuente propietario. No expuso datos de clientes.

Lo que mostró fueron arquitecturas, patrones de diseño, herramientas open-source, decisiones técnicas y lecciones aprendidas. Explicó cómo se construye infraestructura a escala real usando herramientas accesibles. El video es profundamente técnico, profesional y hasta nostálgico.

Lo grave no es lo que mostró. Lo grave es que Atlassian no tenía idea de cuánto dependía de él.

El video no es un acto de venganza. Es una retrospectiva de carrera. Syrakis también habló de crecimiento personal, de cómo aprendió a manejar conflictos, a mentorear, y del desafío de mantener sistemas complejos mantenibles a través del tiempo.

Pero cuando tu retrospectiva de carrera incluye los planos de la infraestructura que sostiene productos usados por millones de personas, estás haciendo algo más que contar tu historia. Estás documentando —públicamente— lo que tu empleador no se molestó en documentar internamente.

Cómo Atlassian manejó el despido

El 11 de marzo de 2026, el CEO Mike Cannon-Brookes envió una nota interna. Las líneas clave:

“Es la decisión correcta para Atlassian. Pero eso no significa que sea fácil.”

“Nuestro enfoque no es ‘IA reemplaza personas’. Pero sería falso pretender que la IA no cambia la mezcla de habilidades que necesitamos.”

Simultáneamente, anunció que el CTO Rajeev Rajan dejaba el cargo y sería reemplazado por dos “talentos de IA de nueva generación”. El sindicato Professionals Australia declaró que los empleados “fueron declarados redundantes sin ser consultados ni recibir ninguna señal de que una reestructuración afectaría sus puestos.”

El paquete de salida fue técnicamente generoso: 16 semanas mínimas de pago, plan de salud extendido, bono proporcional adelantado, $1,000 para tecnología al devolver la laptop corporativa. Slack abierto 6 horas extra para despedidas.

Pero el trato revela la mentalidad: todos son intercambiables. Mismo paquete para 1,600 personas. Mismo proceso. Misma desconexión entre lo que cada persona sabía y lo que la empresa retuvo.

Los 5 errores de liderazgo que hicieron posible el video

No fue un accidente que Syrakis pudiera grabar 38 minutos de infraestructura crítica y publicarlo sin consecuencias legales. Fue el resultado de fallos de liderazgo que se repiten en empresas de todos los tamaños:

1. Despidieron sin evaluar riesgo de conocimiento. En un recorte de 1,600 personas, nadie preguntó caso por caso: “¿qué sabe esta persona que nadie más sabe?” Syrakis no era un empleado genérico. Era el dueño del conocimiento de sistemas que procesan tráfico de millones de usuarios. Nadie lo trató como tal.

2. Trataron a las personas como líneas de Excel. Mismo paquete. Mismo proceso. Mismo mensaje. Una empresa que trata igual a quien construyó el sistema nervioso central que a quien entró hace 8 meses es una empresa que no sabe lo que tiene.

3. No hubo handover ni transición. Del anuncio a la salida: 3 semanas. Sin plan de transferencia de conocimiento. Sin documentación guiada. Sin sesiones de traspaso. El conocimiento se fue con la persona.

4. El timing fue percibido como cínico. Récord de revenue. Ganancias subiendo. Inversores aplaudiendo el recorte — la acción subió 4%. Despedir al 10% de tu gente mientras facturas más que nunca comunica una sola cosa: la gente es un costo que se recorta para invertir en otra cosa. Y cuando comunicas eso, no esperes lealtad post-salida.

5. Reemplazaron al CTO simultáneamente. El líder máximo de tecnología se va al mismo tiempo que 900 ingenieros de R&D. Eso no es estrategia. Es una señal de que ni el liderazgo sabe quién construyó qué.

Lo que un líder inteligente hace diferente

No todas las empresas van a evitar despidos. Pero sí pueden evitar que un empleado con 8 años de conocimiento profundo se vaya y publique la arquitectura entera en YouTube. Aquí hay 6 prácticas que funcionan:

1. Mapa de dependencia de conocimiento. Antes de tocar al equipo, sabes quién sabe qué. No es un documento de RRHH. Es un mapa vivo: ¿qué sistema depende de qué persona? ¿Quién es el único que entiende X? Si la respuesta a esa última pregunta no es “nadie”, tienes un riesgo.

2. Despidos quirúrgicos con handover. Si tienes que despedir a alguien que posee conocimiento único, el paquete de salida debería incluir “X semanas de handover pagado.” La empresa compra el conocimiento antes de que se vaya. Más barato que reconstruirlo desde cero — o que verlo en YouTube.

3. La regla del exit interview técnico. Antes de que alguien se vaya — por despido o renuncia — una sesión técnica donde documenta lo que solo esa persona sabe. No es opcional. Es parte del offboarding. Si no existe algo que solo tú sabes, perfecto. Si existe, se documenta.

4. Cultura de documentación viva. No manuales de 300 páginas que nadie lee. Runbooks. Diagramas. Decisiones de arquitectura con contexto. Mantenidos por el equipo como parte del trabajo, no como proyecto especial que se abandona en el sprint 3.

5. Rotación de ownership. Nadie debería ser dueño único de un sistema crítico por más de 2 años sin un segundo al mando. Si una persona se va y el sistema queda huérfano, el error no es de la persona — es del líder que permitió el punto único de falla.

6. Tratar al que se va como aliado, no como enemigo. Syrakis no salió a destruir Atlassian. Salió a contar su historia. Pudo haber sido embajador de la empresa. Pudo haber sido consultor externo. Pudo haber entrenado a su reemplazo. Lo convirtieron en un whistleblower técnico sin querer, porque lo trataron como un número.

Cómo Guayoyo Tech puede ayudarte

En Guayoyo Tech no damos charlas de liderazgo. Construimos los sistemas que hacen que el conocimiento institucional no dependa de la memoria de tres personas.

Auditoría de conocimiento técnico: Mapeamos qué sistemas dependen de qué personas en tu equipo. Identificamos puntos únicos de falla humana antes de que una salida los exponga.

Implementación de documentación viva: Configuramos sistemas de runbooks, diagramas de arquitectura como código, y wikis que el equipo realmente mantiene — porque las automatizamos dentro del pipeline.

Procesos de offboarding técnico: Diseñamos flujos de handover que protegen el conocimiento: sesiones guiadas de traspaso, documentación de decisiones de arquitectura, y transferencia de ownership antes de la salida.

Rotación de ownership en equipos de ingeniería: Implementamos prácticas para que ningún sistema crítico tenga un solo dueño. Pair programming rotativo, revisiones cruzadas, segundos al mando designados.

No necesitas ser Atlassian para que te pase. Solo necesitas una persona con 5 años de conocimiento único, un líder que no sabe lo que tiene, y un recorte que trata a todos como iguales.

La diferencia entre un embajador y un expositor es cómo lo tratas cuando se va.

En Guayoyo Tech ayudamos a empresas a proteger su conocimiento institucional con sistemas, no con esperanzas.

Agenda una llamada de exploración gratuita. Mapeamos tus riesgos de conocimiento y te decimos exactamente qué proteger, sin compromiso.

The big text on a landing page deserves to arrive. The headline you spent two hours writing shouldn’t just appear — it should choreograph itself into place, letter by letter, so the reader’s eye lands on it and stays.

That’s what @vysmo/text does. It’s a new addition to the vysmo libraries — a tiny, tree-shakable, zero-dependency text animation library with 243 presets and one of the cleanest APIs I’ve used.

This is a tour.

Three lines

The simplest possible use:

import { animateText } from "@vysmo/text";

animateText(element, { preset: "enter/fade-up" });

That’s a real, shippable hero animation. The library splits the text into grapheme-safe slices (more on that below), applies the preset’s choreography, and starts playing on mount. No timeline setup, no .from() / .to() chains, no ref management.

If you’d rather have a different feel, change the preset name:

animateText(element, { preset: "enter/elastic-rise" });   // soft spring
animateText(element, { preset: "enter/bloom-scatter" });  // letters arrive from depth
animateText(element, { preset: "enter/whirl-scatter-curl" }); // dramatic, full-on
animateText(element, { preset: "enter/blur-in" });        // tasteful, agency-deck-ready

Same one line. 120 enter presets to choose from. The catalog at vysmo.com/text has all of them with a live playground — click one, type your text, see it.

243 presets, 3 KB

The headline number is real: 229 generated + 14 curated = 243 presets, in roughly 3 KB gzipped. That’s a smaller bundle than most loading spinners.

The trick is tree-shaking. Each preset is its own ES module export. You import what you ship:

import { animateText, fadeUp } from "@vysmo/text";

animateText(element, { preset: fadeUp });  // pass the object, not the string

When you pass the preset by reference, only that preset’s data lands in your bundle. The other 242 stay tree-shaken out. The string-name form ("enter/fade-up") pulls in a tiny registry — pick whichever ergonomics you prefer.

For a typical landing page using two or three presets, you ship under 2 KB total. For a kinetic-typography portfolio site that uses twenty different presets, you ship maybe 4 KB. Either way, less than a logo SVG.

The presets are split into three categories

Enter (120) — animations for text arriving. fade-up, elastic-rise, bloom-scatter, flip-up-spring, tunnel, plus 115 more. Tuned to feel deliberate, not gimmicky. Most run for ~600–900 ms.

Exit (92) — animations for text leaving. The conceptual mirror of enter, but harder to design well — exits need to feel intentional, not glitchy. fade-down, mist-out, collapse-burst, pinwheel-out. Most react well to being driven faster than their default duration.

Emphasis (31) — animations that loop or play on demand to draw attention. pulse, shake, wobble, coin-flip, spin. These are what you wire to a button click or a “10 items left” badge.

A preset is just data:

import { fadeUp } from "@vysmo/text";

console.log(fadeUp);
// {
//   name: "enter/fade-up",
//   split: "character",
//   stagger: 30,
//   animations: [
//     { prop: "opacity",    from: 0,  to: 1, duration: 500, ease: power2.out },
//     { prop: "translateY", from: 20, to: 0, duration: 500, ease: power2.out },
//   ],
// }

No classes. No inheritance. No registry magic. You can console.log a preset and read it. You can copy one and modify it. You can author your own with the exact same shape.

Two knobs that change everything

Every preset has a default — but the API lets you override any knob per call. Two of them, in particular, are the cheapest way to make the same preset feel completely different.

stagger is the milliseconds between consecutive slices starting. The default is 30 ms (energetic, fast). Bump it to 80 ms and the same preset feels slow and contemplative. Drop it to 10 ms and slices arrive almost together — more of a wash than a choreography.

animateText(element, { preset: "enter/fade-up", stagger: 80 }); // thoughtful
animateText(element, { preset: "enter/fade-up", stagger: 10 }); // urgent

staggerOrder is who gets staggered first. Default is "start" (left-to-right, like reading). But:

animateText(element, { preset: "enter/fade-up", staggerOrder: "end" });    // right-to-left
animateText(element, { preset: "enter/fade-up", staggerOrder: "center" }); // middle outward
animateText(element, { preset: "enter/fade-up", staggerOrder: "edges" }); // both ends inward
animateText(element, { preset: "enter/fade-up", staggerOrder: "random" }); // chaos mode

"center" is the secret weapon. Try it on a hero headline — letters bloom outward from the middle. It feels like the title is being projected rather than typed. Most libraries don’t expose this and people don’t realize how much they’re missing.

Scrolling drives it for free

The handle that animateText returns has a .seek(progress) method that takes a value between 0 and 1 and snaps the animation to that point in its timeline. That’s the only thing you need to drive a text animation from scroll position:

const handle = animateText(headline, {
  preset: "enter/bloom-scatter",
  autoPlay: false,
});

// In a scroll handler — pair with @vysmo/scroll, ScrollTrigger, or your own IntersectionObserver:
function onScroll(progress: number) {
  handle.seek(progress);
}

Cinematic word-by-word reveals tied to scroll position. Two lines. The reason this works cleanly is the library’s single-master-clock architecture — every slice is animated against the same time origin, so seeking is coherent. That’s worth a post on its own (it’s coming).

The split is grapheme-safe

One detail that separates this library from most of its predecessors: the text splitting uses Intl.Segmenter. Which means it correctly handles emoji clusters, regional flags, combining marks, and connected scripts.

// Most libraries:
[..."👨‍👩‍👧"];  // ['👨','‍','👩','‍','👧']  — family emoji shattered into 5 broken pieces

// @vysmo/text:
splitText("👨‍👩‍👧", { mode: "character" }).slices.length;  // 1 — correct

If your animation has ever broken for users in India, Japan, the UAE, or anywhere with an emoji-rich messaging culture, this is why. The fix has been in browsers since 2020 and most libraries still don’t use it. Vysmo does, by default.

Word-mode and line-mode use the same Segmenter, which means they also work correctly in Thai (no spaces between words), Japanese (mixed scripts), and Arabic (right-to-left). If you ship to a global audience by default — which, in 2026, is everyone — this matters.

React, too

There’s a thin companion package, @vysmo/text-react, that wraps the runtime in a declarative component plus a useAnimateText hook:

import { AnimateText } from "@vysmo/text-react";

export function Hero() {
  return (
    <AnimateText as="h1" preset="enter/fade-up">
      Hello world
    </AnimateText>
  );
}

That’s the whole component. Props pass through to the underlying runtime — preset, split, stagger, repeat, all of it.

Try it

pnpm add @vysmo/text

Then three lines:

import { animateText } from "@vysmo/text";
const h1 = document.querySelector("h1")!;
animateText(h1, { preset: "enter/bloom-scatter" });

Open vysmo.com/text for the live catalog and the studio that lets you tweak every knob. Source and issue tracker at github.com/vysmodev/vysmo.

Vysmo libraries are MIT, free, and zero-dependency — there’s no commercial tier, no rate limit, no telemetry. Just one library doing one thing as cleanly as I could get it.

50 crew members. 5 transmissions. Something else on the planet.

Play It First

Space Survival Game

Build a high-stakes, production-grade text-based survival simulation application called “LAST CALL”. ### 1. APPLICATION STRUCTURE & PAGESThe appli

app-boo0vgcv0g01.appmedo.com

Try to get all 50 home. Then come back and read this.

What Is LAST CALL?

LAST CALL is a text-based survival game set in 1974. You play as Lieutenant Kira Malone — acting commander of the ARES-7, stranded on a classified planet called Tartarus somewhere between Saturn and Uranus. Your communications array has power for exactly 5 transmissions before it goes dark permanently.

Earth is 1.4 billion kilometres away. No one knows you’re there.

Every word you type determines who lives.

The Story

In the summer of 1974, NASA launched its most secret mission.

The world was told ARES-7 was an unmanned atmospheric probe. The truth was carried by one man — Mission Commander Harold Voss — in a sealed envelope. The real destination was Tartarus, an undocumented rocky body hidden from all public star charts. The reason it was classified: Tartarus was broadcasting a signal. Repeating. Structured. Deliberate.

50 crew members were told it was a geological survey. None of them knew the truth.

They landed on Day 61. The signal stopped on Day 63.

By Day 66, Commander Voss was unconscious. His sealed orders — missing.

Today is Day 67. You are all that is left.

How It Works

The game runs across five screens:

Setup — Your ship name is randomly generated every run. No two games start the same.

The Story — A classified NASA dossier establishes who you are, why ARES-7 is on Tartarus, and why no one on Earth is coming unless you reach them first.

Mission Protocol — Five rules before you begin. The most important one:

Signal must reach 75% by your final transmission or rescue never locks your position.
Signal below 30% — Tartarus keeps you.

Ship Dossier — Your crew manifest, ship systems reference, last known position. Study it. The AI reads your vocabulary.

5 Transmissions — The entire game. Everything you built toward comes down to what you type here.

What the AI Actually Does

This is the part I am most proud of.

Most AI games use the model to generate story. LAST CALL uses it to read you. Every transmission you send is evaluated across four dimensions simultaneously:

Calm, consistent, signal-focused commands compound across all five rounds. One panicked message in round 3 can undo two good rounds before it. Players feel genuinely sharper after a few runs — not because they learned a system, but because they learned to stay composed under pressure.

That is not a game mechanic. That is a real skill.

The Ship Systems Reference

Not everyone knows 1970s NASA ship terminology. So I built a collapsible Ship Systems Reference panel directly into the transmission screen — always one click away, never in the way.

It tells you what exists on the ship. The reactor core, the signal amplifier, the atmospheric scrubbers, the emergency pods. It gives you the vocabulary to write precise, decisive commands.

It does not tell you what to do. That part is still on you.

The Voyage Report

Every run ends with a full mission debrief — every decision replayed, every consequence revealed, and a Survival Archetype assigned based on your behaviour pattern across all five transmissions.

• Heroic Captain — calm tone maintained throughout
• Cold Rationalist — ship preservation over crew
• Martyr Commander — consistent sacrifice choices
• Humanitarian Leader — crew safety above all
• Strategic Survivor — signal-focused from start to finish
• Adaptive Navigator — inconsistent but resourceful

The report is designed to be screenshotted. What archetype did you get?

The Detail I Am Most Proud Of

On transmission 3, if the crisis takes you to the Tartarus surface — look at the far right of the darkness.

That is all I will say.

What I Learned Building This

MeDo handled this kind of project remarkably well. The multi-turn conversation system carries full context across every transmission — so the AI can reference a choice you made in round 2 when generating round 4’s crisis, without any extra engineering from my side. The game feels alive because the AI genuinely remembers what you did.

The hardest part was not the logic. It was the tone.

Getting the CRT terminal aesthetic right. Making the story feel like a real 1974 NASA document. Writing crisis transmissions that feel urgent without feeling cheap. The game needed to earn its tension, not manufacture it.

I think it does.

Play LAST CALL

Space Survival Game

Build a high-stakes, production-grade text-based survival simulation application called “LAST CALL”. ### 1. APPLICATION STRUCTURE & PAGESThe appli

app-boo0vgcv0g01.appmedo.com

Try to survive. Try to get all 50 home.

And whatever you do — do not let the signal drop below 30%.

Built for the Build with MeDo Hackathon 2026
#BuiltWithMeDo