OpenAI releases GPT-5.6 to select users vetted by US Governmentft.com1 ポイント·投稿者 WithinReason·15 日前·3 コメント
AV2 Codec Looks Like It Will Be Officially Released Next Weekphoronix.com1 ポイント·投稿者 WithinReason·2 か月前·0 コメント
Mozilla to UK regulators: VPNs are essential privacy and security toolsblog.mozilla.org808 ポイント·投稿者 WithinReason·2 か月前·356 コメント
HDMI 2.1 Display Stream Compression (DSC) Ready for Amdgpu Linux Driverphoronix.com66 ポイント·投稿者 WithinReason·2 か月前·49 コメント
Standard Ebooks: Public Domain Day 2026 in Literaturestandardebooks.org382 ポイント·投稿者 WithinReason·6 か月前·55 コメント
Samsung and SK Hynix hesitate to expand production and focus on long-term profitigorslab.de2 ポイント·投稿者 WithinReason·7 か月前·0 コメント
WithinReason·4 日前·議論The planned specs are here, they say it will be made by GL.iNet:https://openwrt.org/voting/2025-02-12-openwrt-twoOtherwise this router from GL.iNet has OpenWRT preinstalled, Wifi 7, 5x2.5G:http://www.gl-inet.com/en-gb/products/gl-be9300
WithinReason·5 日前·議論That's the one that Sam Zeloof is working on, "having lithographically microfabricated various chips in his garage as early as the age of 17"https://en.wikipedia.org/wiki/Sam_Zeloof
WithinReason·6 日前·議論They should have checked if they are more accurate than the consensus, the data doesn't show this.
WithinReason·7 日前·議論Lake Baikal sediment layer almost as deep as the Mariana Trench:https://en.wikipedia.org/wiki/Lake_Baikal#Geography_and_hydr...[...] and below this lies some 7 km (4.3 mi) of sediment, placing the rift floor some 8–11 km (5.0–6.8 mi) below the surface, the deepest continental rift on Earth.
WithinReason·8 日前·議論I'm running Qwen3.6-27B on a single 24GB GPU at 80 tok/s, you don't even need 2 of them
WithinReason·9 日前·議論A general one that covers xcancel:https://einaregilsson.com/redirector/Use this:Redirect: https://x.*to: https://xcancel.$1
WithinReason·9 日前·議論xcancel supports articles now:https://xcancel.com/i/article/2072432348805669139Here is the result:https://github.com/nanocoai/nanoclaw/pull/2742
WithinReason·12 日前·議論> [...] beating Claude Code (32%) at roughly $0.17 per vulnerability foundClaude Code is an agent harness, not an LLM.Claude is a brand (or group of LLMs), not an LLM.
WithinReason·12 日前·議論So a price per GB today is about the same as it was in 2010. 16 year regression, wow!
WithinReason·14 日前·議論So why isn't increasing your velocity from A to B then B to C the same as A to C? Isn't that intuitively transitive too?
WithinReason·16 日前·議論0 because they lack the software, not the HW. The HW works and is relatively easy to make.
WithinReason·16 日前·議論One token per clock cycle at 1B parameters would imply 2 ExaFLOPS, consuming about 10 KWs
WithinReason·16 日前·議論The hardware? Not too difficult, there are dozens of startups. The software? Only NVIDIA could do it so far sufficiently well.