Building the future of secure AI @fdotinc
Joined February 2025
- Tweets 245
- Following 30
- Followers 150
- Likes 7,196
25 Photos and videos
May 12
Another global milestone for Singularity.
We’ve been accepted to Zerobase Season 4 in Seoul.
Zerobase brings together a real Silicon Valley founder network around YC, a16z, Founders Inc., Buildspace, Sam Altman, Khosla Ventures, and more.
From Türkiye to the world.
2
113
We built RoboControl so a fine-tuned Qwen planner can safely control rover fleets.
Voice in. Robot action out.
Every move is checked against telemetry, hazards, policy, and robot health before the rover bridge executes it.
@lablabai @AIatAMD @Alibaba_Qwen
1
2
97
For @AIatAMD , RoboControl includes a Qwen fine-tuning path:
generate rover mission data → fine-tune with LoRA on AMD/ROCm → evaluate JSON validity, safety accuracy, robot selection, action sequence accuracy, and latency → serve with vLLM.
1
2
124
The goal: bring AI agents from text workflows into real-world operations.
RoboControl shows how open models, AMD acceleration, robot security, voice UX, and auditable safety gates can work together to control rover fleets.
2
76
Singularity retweeted
May 7
When raising, remember that you only need one yes.
21/22 VCs passed on the early Anthropic round.
The biggest outcomes are often contrarian, and usually sound ridiculous before the rest of the world catches up.
18
24
434
21,151
The next crypto exploit may not start with a hacked wallet.
It may start with a prompt.
On Solana, one malicious instruction can turn an AI agent into a transaction signer.
Introducing Singularity AIDR for Solana.
Agent Detection & Response for the on-chain AI era.
1
7
164
Singularity retweeted
Apr 29
VC rejection excuses: •
• No moat
• Big labs will kill you
• Too much competition
• Valuation is too high
• TAM is too small
• Get a co-founder
• Not enough traction
• Your retention sucks
Translation: I don’t believe you’ll win.
Your reaction: Watch me.
102
58
1,174
37,031
Apr 24
👏🏻👏🏻👏🏻
Apr 24
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.
🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.
Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today!
📄 Tech Report: huggingface.co/deepseek-ai/D…
🤗 Open Weights: huggingface.co/collections/d…
1/n
1
163
Apr 10
@kanjun We’d love to help with the security side as well. Love the direction you’re taking. Whether now or once it’s open source, we’d be excited to contribute and support you guys.
1
272
🤖 AI agents paying each other in real-time on @StellarOrg
Built StellarAgent, autonomous payment infrastructure with instant settlement, automatic fee splits, and gasless transactions.
Try it live: agenticpayment.vercel.app
2
3
263
Mar 22
I built MicroBox: The fastest & securest open-source sandbox runner. 📦🔒
Tired of rewriting code just to sandbox it? MicroBox runs your UNMODIFIED code safely under an explicit, policy-first environment.
Repo: github.com/SingularityRD/mic…(Thread on benchmarks 👇)
1
1
3
312
Mar 22
How fast is it?
I ran some local benchmarks on a standard home PC. The results: ⏱️ Sequential: 13.17 ms ⏱️ Burst: 18.61 ms
For context, public cloud sandboxes (Daytona, E2B) take 0.20s - 3.9s for provisioning first command. Local policy execution is blazingly fast.
1
1
128
Mar 22
What ships today?
✅ Safe by default, explicit policies
✅ Linux secure backend, macOS/Windows compat
✅ Built-in validate, doctor & bench commands
My next goal: Building an integration so you can execute code directly via @OpenAI @OpenAIDevs Codex.
1
107