Hey Product Hunt — maker here. I built The 8088 because the AI news firehose was breaking my brain. Every morning was 50+ tabs of half-skimmed tweets, blog posts, and press releases — and I still felt behind. I wanted three things no aggregator I tried gave me: 1. Real curation. Not "more articles" — significance scoring. Did this matter? Rated by an LLM, not a popularity heuristic. 2. More than headlines. Which companies and people are showing up? Which categories are spiking? Which predictions came true? Same article enrichment, five different views over one DB. 3. No cloud LLM bill. Every summary, score, sentiment read, and entity extraction is done by Gemma 4 26B running on a server in my home (vLLM, nightly systemd timer, static site rsync'd to shared hosting). Marginal cost per article: electricity. How the approach evolved: started as a single feed page. Then I realized the same enrichment data could power totally different products — so I added entity tracking, then a StockPulse view, then a "what are McKinsey/BCG/Deloitte writing about AI" tracker, then a fantasy league where you draft AI companies and compete weekly on whose news category gets the most coverage. Half-joke, half-real. It's named after the Intel 8088 — the chipset on the first PC I used. The "running it locally" thread is the throughline. Would love feedback on what's useful, what's missing, and what to kill. Especially curious what people think of the significance score — that's the hardest part to get right.

The 8088

Curated AI news, ranked by a local LLM (no cloud)

Curated AI news, ranked by a local LLM (no cloud)