Trending Wire
The loudest threads, videos and posts about tech right now, ranked by how fast they are climbing rather than how big they already are. Each one links to the original.
One thing I’ve always wanted in more apps, so I added it to # Cue : Move tools and navigation to where they work best for you — bottom, left, or right. Different screen sizes, different ways of holding a device, different workflows. The most useful controls shouldn’t have to live in the same place for everyone. # iOS # automationtools # buildinpublic # indiedev # productivity # accessibility
OPPO F35 & F35 Pro Unboxing & Test📌 1.5K AMOLED, 4K Ultra-Clear Video, 8000mAh & more
Dont Get iPhone 18 Pro | 10 days Review | Malayalam
Get rid of this crap. This will backfire on you. And cause outrage in the community. It will force me to buy off of Steam with Steam keys if sales differ for each person. 1/2 #steam #steamnews #gaming #pcgaming #pcmr #pc steamcommunity.com/groups/steam... Steam :: Steamworks Developmen...
Fuck Vanity Fair *and* Annie Leibovitz https:// bsky.app/profile/shannonvallor .bsky.social/post/3mx5z3f3s6c2k # AI # LLM # SamAltman # OpenAI
Riot anticheat head Phillip Koskinas asserts that the company's hardware bans last four months, and cannot "infect" other components when inserted into a new system. pc-gamer.visitlink.me/0CHoUp #pcgamer #pcgaming #riotgames #riot #valorant
How is it possible that qwen 27b is so good? When GPT 4o had a trillion parameters and was worse?
That's my senator. # AI https://www. civilbeat.org/2026/10/schatz-p ushes-ai-oversight-as-tech-goes-rogue-humans-in-control/
🎮 Supermassive Lays Off 70 People Less Than Six Months After Releasing Its New Sci-Fi Horror Game Directive 8020 was met with mixed reviews on Steam 📰 Source: Kotaku 🔗 Link: https://kotaku.com/supermassive-lays-off-70-people-less-than-six-months-after-releasing-its-new-sci-fi-horror-game-2000740078 # SteamDeck # PCGaming
When Redditors come in here and ask why we run LLMs, this is why: Big AI is watching.
Make no mistake, selling 64 GB DGX Spark variants at the same cost as the original 128 GB is straight drug dealer behavior.
llama.cpp v0.6.0 released with MTP speculative decoding for Qwen4Exp and lots more
Smallest Jev-like model
Whistle: speech to text in a 16.9MB file
Qwen image 2.1 newer than Qwen image 3.0?
Update #5: Post training yandex/AliceAI-80B-A3B [instruct!] from scratch
Why is Qwen3.8 refusing to write any actual code locally? It's just going round and round repeating design decisions, structure and a bunch of stuff. It will write a bit of code on screen but not to the project folder. Maddening!
Update to my current rig
While I was investigating why my opencode context was large I created the opencoder-leaner to try prune the context (minimal prune in the tools, agent pi-like, bash only)
I ran 1,200+ Blender modeling runs across LLMs and agent harnesses and made them votable
What comes close to Codex's Computer Use MCP
Blackwell + consumer GPU
Agens Volundr 32B Preview: our small team's first model on our own hybrid architecture. Only 18 of 72 layers keep a KV cache (Apache-2.0)
nokia-applied-research/AnyJev: Turn any LLM into a Jev-style decision model
We swapped AdamW's optimizer states for a Fast Fourier Transform (FFT) to cut VRAM in half. Anyone else trying non-quantization methods?
Agent: Muse, but open source and living on your Android phone
Y'all this is a sexy paper; context language models
How to use Local Models to monitor your screen. Open Source, No Install and Completely Free!!
Qwen3.8-Flash-Next (125B) on a single Strix Halo mini PC: 44-59 tok/s with speculative decoding, ~1,400 tok/s prefill, engine is open
Just joined the local LLM club! What's the best way to stay in the loop on the best local models?