Elvira D (@elvirafortune) header
Elvira D (@elvirafortune) profile picture

co-founder @fastshot_ai (YC F25), ex @google

San Jose, CA
Joined March 2016
Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)Photo by Elvira D (@elvirafortune)
This is unbelievable 😡
If you've wondered where I've been recently, I gave an interview. This is a very difficult time for me. I'm trying to bring my mom home, and I could use your help spreading the word. Please watch, and please share. interview: youtube.com/watch?v=JlK2a6yv… website: bringmymomhome.com/
7
238
Finally someone is killing inefficient sales
Announcing @LeadbayAI $4.3M to kill everything you're doing with sales. Corgi, Deel, L'Oréal recently told us: "We found in 10 min on Leadbay what we couldn't in 6 months with Clay and a GTM agency." Leadbay qualifies the millions of data-signal-scarce companies nobody could find before. This is how it works:
6
233
That's the agi of film making
Today we're launching FilmArena [dot] ai. Biggest release we've ever shipped. The problem: a new SOTA AI image & video model drops every week. @OpenAI shipped GPT Image 2 literally yesterday. 1000 endpoints, no honest answer to "which one is best right now?" Arena AI and Artificial Analysis cracked LLM benchmarking. But for creative AI, they're solving the wrong problem. There is no "best image-to-video model." Best at what? World building? Action Sequences? Expressions? Every category has a different winner, and it flips weekly. So we benchmarked 40+ SOTA models with 2M+ votes across 190 countries. By subcategory. Re-ranked every week as new models ship. With thousands of real humans. But that's just half of it. Then there's the SANDBOX. Higgsfield and other aggregators put models in one place. But creators still juggle twelve subscriptions to do one job. FilmArena collapses it into one tab. Every frontier model. One interface. Same price. Call the best, or call 3 at once and compare. The AI film stack has been fragmented since day one. Today it isn't. Go try it free: filmarena [dot] ai Already wild results from this week: 1. @OpenAI GPT Image 2 (released yesterday) dominates image editing, but @GoogleDeepMind Nano Banana 2 still wins animated 2. @pixverse C-1 beats @dreamina_ai Seedance 2 on animated 3. Veo 3.1 Fast outperforms Veo 3.1 on realistic video And a lot more. Go check it out.
1
3
233
As always, some scheming tactics... to generate more revenue from the masses 🌶️
Apple has commented on the reasons behind Cal AI removal for TechCrunch and it wasn’t about external payments.​​​​​​​​​​​​​​​​ There were multiple violations: First, the app removed Apple IAP as a payment option, replacing it with an embedded Stripe flow, a violation of Guideline 3.1.1, which requires IAP to be offered alongside any external link. Second, the paywall was designed to mislead users: the weekly price was displayed more prominently than the actual amount billed, a violation of Guideline 3.1.2c. Third, users who declined the first paywall were shown a second offer with different terms. This triggered a wave of negative reviews accusing the app of being a scam. After the developers made the necessary changes, Apple reinstated the app. A telling story: remove the app and explain why. A clear message to developers about what is not allowed. Shortly after Cal AI returned, Apple also removed many copycat apps.
5
174
Guess what Forbes 30 under 30 🍋🍋🍋
This was Cal AI's policy breaking "second offer"
5
131
New era of mobile creation!
We got into Y Combinator to kill AI slop for mobile apps. 15,000+ users. 16,000+ projects. 96 million lines of code later, we’re finally ready to introduce Fastshot. The team: → @DmitryFatkhi was building agents for @GoogleDeepMind teams at @Google → @elvirafortune was part of the team that created @kubeflow and built AI developer tools at @Google → @alexndrsukhanov scaled streaming for security cameras at @meraki We're the only AI mobile app builder that aligns with you at the design stage — before anything gets built. Describe → Design → Build → Ship. No more vibe-coded looking apps. No more wasting credits fixing ugly visuals. All in one platform. Welcome to how the future of apps should be built.
3
4
267
🚨 you don't want to miss this
If you had 3 hours to build something that could raise $5M… This is your shot. Stanford x DeepMind x Fastshot Hackathon — April 12 We’re bringing together top builders, cutting-edge AI tools, and VCs who are actively writing checks. ⚡ Build a web or mobile prototype in just 3 hours   🧠 Use @GoogleAIStudio + @fastshot_ai  💰 Pitch directly to investors writing $500K–$5M checks   🏆 What you win: • 30-min pitch meetings with top-tier VCs   • $70K+ in prizes ($60K IRL + $10K online)   • Serious exposure to funds that matter   👀 Judges & investors from: Investors & Angels: @AnithaVadavatha (AB Plus Ventures) @CCgong (Menlo Ventures) @DanielDarling (Focal VC) @denwezoh (Susa VC) @gstuto (Offscript VC) @omri_drory (NFX Bio) @ShahramSN (Civilization Ventures) @vigsachi (Gradient Ventures) @francescagnuva (Audacious Ventures) @i_arp1t (Threshold VC) @NancyZWang (Felicis) @prateekj (Moxxie Ventures) @brianzhan1 (Striker Venture Partners) and many more. Google / DeepMind: @jocarrasqueira (Product Lead, DeepMind) @ivanleomk (DevEx, DeepMind) Amit Vadi (Head of Community, DeepMind) Stanford: Prof. Nikesh Kotecha (Head of Data, Stanford Health) @tagrtagr (VP Startup Development, Stanford Bases) Co-organizers: @elvirafortune, @dmitryfatkhi, Jesse Z., Arpiné Arakelyan, Kate Polet Approval required. Spots are limited.
1
8
435
🚨 BREAKING: Elite hackathon with a chance to secure up to $5M in seed funding*! Top teams will be judged by representatives from Stanford, Google, and large venture capital funds. Win direct 30-minute pitch meetings with top VCs looking to write $500K–$5M seed checks, plus a share of a massive prize purse ($60K for the In-Person track + $10K for the Online track). RSVP link below 👇
4
4
10
707
🚨 over 94% of YC W26 companies build AI products
4
229
OpenAI bought this GitHub repo. 127 Fortune 500 companies were already using it. It is still free. I found a repo called promptfoo a few weeks ago and I wish I found it earlier. It is an open-source tool for testing LLM prompts the same way you test software. You write test cases in YAML and define what the output should look like. Then you run it against any model and get a pass-or-fail report with side-by-side comparison. It supports 77 providers including OpenAI, Anthropic, Gemini, Bedrock, and local models through Ollama. It has 66 types of checks, from simple ones like "output contains X" to harder ones like "is the answer factually correct" scored by another model. Here is why this matters. The Vectara Hallucination Leaderboard tested 98 models across 7,700 articles. The best model hallucinated 0.7% of the time and the worst hit 30%. Most models sit somewhere between 1% and 10%. And the same model can give different answers to the same prompt on different runs. If you swap a model, change a system prompt, or update your context window, your outputs can break in ways you will not notice until a user complains. Prompt testing catches this before it ships. The numbers behind promptfoo are real. Over 350,000 developers use it and 127 Fortune 500 companies run it. OpenAI acquired the company in March 2026 and kept it open source. It also has a red-teaming engine with 72 vulnerability plugins that test for prompt injection, data leaks, and jailbreaks. We have a similar system but it is built for our specific use case. Our platform generates working apps from prompts, and we control the whole proccess. From design to production app maintainance. So we have 20+ agents running each can use different models case by case. Each model needs its own prompt because they all have different strengths and weak spots. But for general LLM work, promptfoo is the best tool I have seen. It runs locally, your prompts stay private, and you can plug it into CI so tests run on every change. If you ship AI features without testing your prompts, you are shipping untested code with extra steps.
5
229
As a founder and CPO I saw it coming and never prioritized it for @fastshot_ai Soon we will see many vibecode apps banned on App Store.
Apple Quietly Blocks Updates for Popular 'Vibe Coding' Apps macrumors.com/2026/03/18/app…
1
2
317
Do you remember when you joined X? I do! #MyXAnniversary
1
1
114
🚨This mistake slows down your AI (according to a recent paper) Your CLAUDE.md file is loading stale data into every single AI session. Here's how to fix it in 5 minutes. The official docs say run /init to generate it. I used to say the same. Then I read a paper: "Evaluating Agents.md: Are repository-level context files really helping code agents?" (paper link in comments). The answer: often, no. Here's why. Claude Code reads your CLAUDE.md on every single message. Every session. If that file has outdated info - old folder names, commands that changed, things the AI could just look up itself - you're giving it wrong data every time. Think of context like RAM. You have 1 000 000 tokens. Claude Code's own setup takes about 2%. Then your CLAUDE.md loads. If it's 500 lines from /init, you wasted a lot of space on info the model could find itself in 5 seconds. Real example: your CLAUDE.md has a folder path. You rename the folder. Now the file is wrong. Claude Code follows broken instructions. That's worse than no instructions at all. So what should be in claude.md? Only things the AI can't figure out on its own. Your rules, not your project structure. Like: - "Name all files in kebab-case." - "Don't run npm run build unless I ask." That's it. Two lines is enough for most projects. Everything else - docs, schemas, deployment guides - put it in a /docs folder. The AI won't load it automatically. You share it when you need it. For things you do often? Use skills. Small markdown files in a /skills folder. Claude loads them when the task needs it. Not on every single message. Less context, better results. It adds up.
4
196
Didn't get into YC, consider this 👉
YC Startup School is back this summer! > 2 days, in-person with the most ambitious technical builders in the world > Talks from Jensen Huang, Sam Altman, Alexandr Wang, Jeff Dean + more > Small group sessions with YC partners > $25k+ compute credits > Hands-on hardware/robotics demos > Research poster sessions
5
261
The future belongs to people who can go from idea → live app fast. That’s what this sprint is about. Let’s build.
YOUR IDEA DESERVES MORE THAN NOTES. Too many great ideas die in Notes apps or in “I’ll do it later.” Time to turn your idea into something real. We’re organizing a one-month challenge with @fastshot_ai where you’ll build an app from scratch and no technical experience needed.
1
9
328
$5K in prizes. 30 days. Let’s see what you’ve got. 👉
We just partnered with @fastshot_ai for The Ultimate Mobile App Sprint. $5,000 in prizes. One month. Build a real mobile app that makes money, no code needed. Details: → March 2 to April 2 → Build with @fastshot_ai (AI mobile app builder, YC-backed) → Monetization handled by @adapty → $2K first place, $1.5K for 2nd and 3rd → Use code "SPRINT" for Fastshot Pro at $1/month Anyone can enter. Register below (link in comments) The NCF team is entering and building in public. But we need you to pick what we build. A few ideas 👇 🔍 AI Niche Finder 🛠️ No-Code Stack Recommender 📋 Founder Daily OS 💡 Side Hustle Validator Which should we build? We'll announce our build choice soon. Follow along. Or better yet, enter alongside us. Sign-up below 👇 #UltimateAppSprint
3
135
Are the days of disguised scammy marketing finally over?
Today we're announcing Paid Partnership labels on posts. X's core value is providing on authentic pulse on humanity. While we want to encourage people to build their businesses on X, undisclosed promotions hurt the integrity of the product and lead people to distrust the content they read on X. This new feature will allow you to comply with regulations, but more importantly: it enables you to be transparent with your followers.
2
166
This ad is controversial. But honestly? If you watch it with real empathy, it hits deep. I’ll take raw, real, unfiltered stories over polished, overproduced, fake-perfect demos any day.
2
193