Recent submissions

Comment · 11 months ago
YC W25 Biotech Startup Hiring – Build Programmable Therapeutic Vines

We’re EPIVINE—engineering self-replicating vines that deliver epigenetic age-reversal therapies as over-the-counter botanical supplements, bypassing pharma patents. Backed by YC, we’re turning plants into living pharmacies.

We’re hiring:

1. Senior Synthetic Biologist – Engineer dCas9-p300 nanostructures & optimize epigenetic editing. 2. Biomaterials Engineer – Design shape-memory alloy scaffolds that guide vine growth. 3. Computational Biologist – Model aging reversal (H3K9me3) and analyze epigenetic beacon data. 4. Full-Stack Biotech Dev – Build the IoT and data platform for our smart vineyards. 5. Head of Regulatory Strategy – Navigate FDA botanical supplement pathways and secure OTC status.

Why join? Founder-market fit (published epigenetic plant research), massive aging/wellness market, 3 provisional patents, and YC network. Located in South San Francisco.

*Apply:* Don’t send a resume. Pick a role, build something relevant in 48 hours, and email it to [email protected] the subject: `[Role] – [Your Name] – 48hr Build`.

We’re building the future where your garden treats disease. Help us grow it.

Comment · 1 year ago
Recently, I have been working on a Digital landfill viz. Flink job: DePIN Resource matcher. A real time matching engine that connects resources offers with requests. Source: Kafka topic(beacon.resource.offers: Beacons advertising their available Npu compute for ambient LLM inference). Source 2: Kafka topic inference.requests(Requests for LLM inference that needs resources).
Comment · 1 year ago
Defect detectives - The giga yield case files. My PTG spots 99.9% of giga flaws - Solve a case, win $25k or a live NFT dashboard pass. Each week, a "crime scene" dataset is dropped from a simulated Tesla Giga factory line, Data is 95% sparse, mimicking real-world sensor noise and incomplete logs. Players must detect "defects" ( e.g., micro-anode cracks,electrode delamination, thermal hotspot anomalies) Gameplay loop is simple 1. Players join via X threads or 'n "case assigned. 2. Download Giga dataset fragment ( Structural CSV/API feed) 3. Runs analysis with PTG ( Probablistic Tesla Graph) or custom script 4. Submit predicted flaw coordinates.
Comment · 1 year ago
IFS Bottleneck means; NVIDIA buys time, Intel loses it and Blackwell ultra slips. RISC-V (Tenstorrent) + AMD ROCm keep the pipeline moving while NVIDIA polishes it's hammer. Lesson; Adage (Old Norse inspired)

“Hamarr klofnaði, er vé fann eigi leið; Spjót fló, en skjár var eigi bundinn.”

Translation back to English

“The hammer split, when craft found not its path; The spear flew forth, yet screen was never bound.”

This mirrors the original iambic pentameter adage, but cast in skaldic-styled compactness, tying the hammer (Mjölnir) to the broken driver/kernel path, and the spear (Gungnir) to the unpatched Wayland/X11 display stack. Something like Snorri Sturlosson might've slipped into the prose Edda? except it's about GPUs and Linux drivers.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton