Rohan Mayya

We released Unleash the Avatar's pre-alpha gameplay trailer yesterday, and we've already gone viral on Bilibili in China (over 2.5 million views on the IGN video alone) Thought I'd use this thread to summarize some of what we've done in the past 4 months:

Hiring form link: binary.so/Dp1T8NC

Hiring: Environment Artist Looking for someone who wants to craft high quality, Indian looking environments in Unreal Engine (screenshot is from our game) alongside a strong team In-person only, in our office in Bangalore Apply through the form below!

As I said, they're not bottlenecks for teams that think through them We're not expected to do live physics sims when they can be baked, we use cheap methods to fake cloth and other sims (like RBANs in Unreal) etc The point I'm trying to make is: pixel shaders can get out of control even if you try your best to optimize it, because of modern gaming graphics demands The others are all controllable with good decisions

@rohan_mayya Physics,AI, etc are huge bottlenecks in games, especially open world or simulation heavy ones. Draw calls are also one of the best (and trickiest) things to optimize on the CPU side - and they directly impact how well the engine can feed the GPU.

Something I've noticed when people talk about games performance is a huge focus on things like polycount, or texture size, or lighting (GI vs Baked vs Light Probes, etc) - but not a lot of focus is put on things like AI, Pathfinding, particles, pooling, physics, skinned meshes, etc. Is it because that stuff is just... harder to see? To understand? Or is it too 'disconnected' from the more visual stuff? Honestly the amount of geometry that engines can crunch through now - even without things like nanite or virtual textures - is incredible. I've found that the real bottlenecks aren't graphics - its everything else.

I have never met someone high ELO who's a bad hire Fun fact: we have someone on our team who is top 100 SEA on Deadlock!

A silent pandemic is underway in Indian cities. Far too many couples are opting to not have kids at all. It is the wrong decision. Most will regret it.

2 weeks to go

Many people trying to cancel AI progress saying "why are the labs promising a cure for cancer but taking away my white collar job instead" should know that innovation is really messy Sometimes better understanding in one domain indirectly translates to another, which is you can't halt progress on any one axis Examples: - GPUs were originally made for games/3d graphics, but now power modern deep learning, - ARPANET began as a research network, but became the foundation for the internet, - TEFLON started in refrigerant research, but now used in nonstick cookware, - Game engines for games, but now used for archviz, virtual production, automotive sims etc

Hiring for the One Job AI Can't Take. This post is the screening round itself. If you read it honestly, test yourself against it, and recognize yourself in it, you have already cleared round one. The exercise at the end is for round two. Part 1: The Great Reallocation We are living through the largest reallocation of human labor in recorded history. AI is making the data-heavy parts of knowledge work radically cheaper, and the standard story is that this displaces humans. We think the opposite is closer to the truth. The net result is not pure job loss but a reallocation of human labor toward the things that make us most human: tacit judgment, taste, trust, social skill, embodied action, a sense of purpose and meaning. We call this “The 100x Dual Paradox”. We are witnessing the biggest change in the history of education, employment and entrepreneurship. Part 2: What We Are Actually Building At 100xEngineers we prepare builders for this shift, and we intend to go further. We are not trying to run education programs. We are trying to build the world's largest community of high-agency builders, with its own culture and its own economy. We have trained over fifty thousand professionals so far. That is the starting condition, not the achievement. The real work is what those professionals become to each other, and what this ecosystem becomes, over the next twenty years. Our species runs on shared stories, and building this community is, at its core, the work of making a new shared fiction real. The role we're hiring for is the person who holds that fiction and makes it livable for the next generation of builders. It is probably the most ambitious project of our working lives. Part 3: The Person We Are Looking For We want to be careful with the words, because "Community Lead" is a title that has been diluted by an entire industry of engagement managers, content calendars, Discord moderators, and growth hackers. That is not what this is. A traditional community manager optimizes for activity. This role exists to build belonging, shape culture, and influence the future of human communities. So here is what we are actually looking for. Someone who is already, in their own life, quietly obsessed with a particular set of questions. What makes a group of people unite across decades? Why some communities calcify and others renew themselves? What's the difference between an event, a movement, and an institution? How small rituals compound into shared identity? How a group remembers its own history? If these questions are new to you, this role might not be a good fit for you. If you have been thinking about them for years without anyone asking you to, we should talk. Part 4: What This Role Is Not We want to be equally explicit about what we're not looking for. Not someone whose instinct is to run the community like a growth funnel. If your proudest achievement is a 3x lift in weekly active users, we are probably not a fit. Not performative warmth. The exclamation marks, the "so excited to welcome everyone" energy, the manufactured enthusiasm every SaaS community has adopted. The members we attract are high-agency builders, and they can tell when warmth is real and when it is a brand voice. Not someone who sees this as a content calendar job. Posts, threads, AMAs are downstream of culture, not upstream. Not neutrality. 100x has a worldview, and the person who holds our community should be able to articulate it, embody it, and defend it publicly. Part 5: Location & Practical Notes This role is based in Bengaluru and is fully onsite. The culture of 100x is built in rooms together, in the texture of how we argue and build and gather. We do not think this role can be done well remotely. If Bengaluru does not work for you, this is not the moment for us. The best person for this role may not come from a traditional community management path. They might come from teaching, editing a magazine, organizing a movement, running a meaningful offline gathering for years, studying religion, history, or anthropology. From being a writer whose real interest is institutions. A decade of platform experience at a SaaS company may actually be a negative signal, if it has calcified a particular way of thinking about what a community is for. You will work closely with me and the founding team. You will have the latitude to shape the culture, the rituals, the language, and the rhythm of one of the largest builder communities in Asia for the next decade. The Exercise If you have read this far and felt recognized rather than exhausted, here is what to do. Write down three things. 1. A community, movement, or gathering you have been part of or built, that meant something to the people inside it. Tell me in one paragraph what it actually felt like to be a member, not what the KPIs were. If it no longer exists, tell me why it ended. 2. One concrete weakness you see in 100xEngineer Community today, from the outside. Something you think we are getting wrong, or missing, or under-investing in. 3. One ritual, tradition, or piece of shared language you would introduce in your first ninety days, and why that one. Keep it short. If any of these questions made you put the tab away, that is useful information for both of us. If they made you open a blank document, Send it to us when you are done.

Got a chance to judge a hackathon hosted by @crustdata and yc last weekend thanks @ManmohitG and @TheChowdhary for the invite, and congrats on the insane growth of the company!

A suggestion I would give to my younger self to generate better ideas: shadow a company in a specific industry I know many smart engineers who find it hard to generate ideas (so they gravitate towards note takers, generic phone/b2b apps, personal assistants etc) Spend 1 month with us at our game studio, and you will 100% walk away with 5+ ideas that we would pay for immediately Maybe I should compile a list?

To summarize.. - bloated teams and diluted ownership, - 5-8 year dev cycles with multi-million dollar budgets (sometimes exceeding $100m), - overhiring, and then layoffs after every launch, - push agenda over good gameplay and story

Work like the company is your own

안녕하세요

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

Introducing Atlas 1. Willow's new frontier speech-to-text model. It outperforms ElevenLabs, Deepgram, OpenAI, and more by a wide margin. Built on the first scalable, human-powered transcription infrastructure ever built for real-time dictation.

@WillowVoiceAI Congrats!!

How are so many people engaging with this slop? This info is completely false and the post style itself is teetering on borderline LLM psychosis

Looks like figma finally caught up with @paper!

I’m so excited for the next Unleash the Avatar trailer. Homegrown engineering, animation, and art excellence. Maybe we’ll do a a playable roadshow side by side too so many of you can experience it. Spent a considerable amount of the last 18 months of our life on this.