Ace: aceofspadeshq at gee mail.com
Buck: buck.throckmorton at protonmail.com
CBD: cbd at cutjibnewsletter.com
joe mannix: mannix2024 at proton.me
MisHum: petmorons at gee mail.com
J.J. Sefton: sefton at cutjibnewsletter.com
Chavez the Hugo 2020
Ibguy 2020
Rickl 2019
Joffen 2014
AoSHQ Writers Group
A site for members of the Horde to post their stories seeking beta readers, editing help, brainstorming, and story ideas. Also to share links to potential publishing outlets, writing help sites, and videos posting tips to get published.
Contact OrangeEnt for info: maildrop62 at proton dot me
A full announcement is expected at CES and general availability by the end of the January.
It includes new performance and efficiency cores, a low power core that isn't described in detail yet, and a new Xe3 graphics engine that provides a preview of the upcoming Celestial discrete GPUs.
Also, while Lunar Lake was one size fits all, there are three different versions of Panther Lake:
The cheapest version has four P cores and four LP cores, four Xe3 graphics cores, supports soldered or socketed RAM, and has eight PCIe 4.0 and four PCIe 5.0 lanes.
The version for high-end laptops with dedicated graphics has four P cores, eight E cores, four LP cores, four Xe3 graphics cores, supports soldered or socketed RAM, and has eight PCIe 4.0 and twelve PCIe 5.0 lanes.
And finally a model for fast laptops without dedicated graphics has four P cores, eight E cores, four LP cores, twelve Xe3 graphics cores, supports soldered RAM only, and has eight PCIe 4.0 and four PCIe 5.0 lanes.
So you has to pick and choose. There is no model that gives you everything.
Though potentially the last of these can support CAMM2 modules to allow for expandable LPDDR5X RAM, and four lanes of PCIe 5.0 is enough to drive a mid-range laptop GPU.
It's also supposed to user 30% less power than Lunar Lake thanks to the move from a 3nm to 1.8nm process.
According to a study by Anthropic, it is remarkably easy to craft malicious data that will turn any LLM into a train wreck. Testing on Meta's Llama and GPT 3.5, as well as the open source LLM Pythia, this worked 100% of the time.
What's more, larger models offered no increase in protection. The poison pill of 250 documents worked on 13 billion parameter models just as it did on 600 million parameter models.
Radiance cores (for enhanced ray tracing support), neural arrays (for AI), and universal compression will all be coming to RDNA5 and, presumably, the PlayStation 6 - because AMD designs the chip inside the PlayStation.
Three full-size Zen 6 cores, eight smaller Zen 6c cores (about 75% of the clock speed), 68 RDNA5 cores, and 48GB of GDDR7 RAM on a 192-bit bus.
48GB of RAM will change things for developers, and the jump from Zen 2 to Zen 6 will speed things up. Moving from 52 RDNA2 cores to 68 RDNA5 cores will provide... Twice the graphics performance? Something like that.
And yes: AMD designs the chip inside the Xbox as well.