Veteran Windows dev shows off AI running on 47-year-old PDP11 with 6 MHz CPU and 64KB of RAM
The suitcase-bound robot is a decent performer, too. A roughly 200 ms Time To First Token (TTFT) means Sparky can start formulating responses very fast, and then runs at about 14-15 tokens per second, according to the LLM enthusiast. What’s more, the response is natural (for a robot), using SenseVoiceSmall for speech-to-text and Piper for text-to-speech. Piper is synced with the PixiJS face, and the mouth animation updates at 43 Hz. CreativelyBankrupt notes that “Vision and OCR are native to Gemma 4 now.” You can also configure and interact with Sparky using a button row, a joystick, and an analog encoder knob.
CreativelyBankrupt shared the genesis of Sparky on Reddit, partly to seek hints and tips on optimizing Gemma 4 E4B on Orin-class hardware . However, respondents are mostly in awe of this “cool… adorable… cute… kitschy…” project.
Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/maker-packs-an-opinionated-googly-eyed-ai-chatbot-into-a-mobile-suitcase-powered-by-an-nvidia-jetson-entirely-local-machine-entity-runs-gemma-4-e4b-and-can-respond-in-200ms#main
- https://www.tomshardware.com
- Samsung starts winding down chip production six days before planned 18-day strike — company enters 'emergency management mode,' daily losses could hit $2 billio
- TP-Link Archer BE770 Wi-Fi 7 router review: Prioritizing wireless performance over wired connectivity
- China bypasses US GPU bans with 1.54-exaflops 'LineShine' supercomputer — CPU-only monster packs 2.4 million Huawei-designed Armv9 cores
- Samsung meeting transcripts show memory workers offered incredible 607% bonus worth $477,000, while logic chip staff get as little as 50% — union says misbalanc
- China bypasses US GPU bans with 1.54-exaflops 'LineShine' supercomputer — CPU-only monster packs 2.4 million Huawei-designed Armv9 cores
Informational only. No financial advice. Do your own research.