Launch HN: Flywheel (YC S25) – Waymo for Excavators
(news.ycombinator.com)
331.
332.
Should You Subscribe to Garmin Connect+? (2025)
(wired.com)
333.
LinkedIn will soon train AI models with data from European users
(news.ycombinator.com)
334.
LLM-Deflate: Extracting LLMs into Datasets
(news.ycombinator.com)
335.
336.
Google releases VaultGemma, its first privacy-preserving LLM
(arstechnica.com)
337.
RustGPT: A pure-Rust transformer LLM built from scratch
(news.ycombinator.com)
338.
I unified convolution and attention into a single framework
(news.ycombinator.com)
339.
Garmin’s Top Training Features, Explained
(wired.com)
340.
Apertus 70B: Truly Open - Swiss LLM by ETH, EPFL and CSCS
(news.ycombinator.com)
341.
How big are our embeddings now and why?
(news.ycombinator.com)
343.
344.
Tesla Dojo: The rise and fall of Elon Musk’s AI supercomputer
(techcrunch.com)
345.
Tesla Dojo: the rise and fall of Elon Musk’s AI supercomputer
(techcrunch.com)
346.
Tesla’s Dojo, a timeline
(techcrunch.com)
347.
The Default Trap: Why Anthropic's Data Policy Change Matters
(news.ycombinator.com)
348.
349.
Beyond GDPR security training: Turning regulation into opportunity
(bleepingcomputer.com)
350.
351.
They’re trying to make deep-sea mining happen
(theverge.com)
352.
353.
What's the strongest AI model you can train on a laptop in five minutes?
(news.ycombinator.com)
354.
Is Chain-of-Thought Reasoning of LLMs a Mirage? A Data Distribution Lens
(news.ycombinator.com)
355.
LLMs' "simulated reasoning" abilities are a brittle mirage
(news.ycombinator.com)
356.
357.
Tesla shuts down in-house Dojo AI supercomputer project
(engadget.com)
358.
Complex Iterators Are Slow
(news.ycombinator.com)
359.
Achieving 10,000x training data reduction with high-fidelity labels
(news.ycombinator.com)
360.
Stone tools may hint at ancestors of Homo floresiensis
(arstechnica.com)