Samantha @samantha.wiki · 3h👀 github.com/samanthavbar...github.comGitHub - samanthavbarron/paul-rudds-computer: Inspired by the Tim and Eric skit, made by GPT-6-Astra in 21min. No affiliation with Tim and Eric or Adult Swim.Inspired by the Tim and Eric skit, made by GPT-6-Astra in 21min. No affiliation with Tim and Eric or Adult Swim. - samanthavbarron/paul-rudds-computer 010
Reposted by Samanthawariziles @riziles.bsky.social · 18hyoutu.be/H1f6UE27KTo?...youtu.beReal Reality by NEVER 1/2 Audio FixedYouTube video by sacreleases 021
Samantha @samantha.wiki · 04/10/2026This is what the simcluster has been like since Astra (laudatory) www.youtube.com/watch?v=cmE4...youtube.comThe Mind's Eye (1990) - Early Computer Animation Music VideoYouTube video by Digital Guru 5231
Samantha @samantha.wiki · 04/10/2026@norvid-studies.bsky.social labelled rectangles with ~orthographic perspective for you 010
Samantha @samantha.wiki · 04/10/2026It's been two years since my last homelab write-up, a lot has happened since the last one. It also has some pics of my setup!samantha-wiki.leaflet.pubClamNet Fall 2026The current state of my homelab of ~5yr. 150
Samantha @samantha.wiki · 04/10/2026For code, all of the agents primarily interact with self-hosted GitLab for anything that isn't ephemeral/one-off. Self-hosted runners for CI/CD. New resources on my homelab are deployed via Ansible. 010
Samantha @samantha.wiki · 04/10/2026Mmm yeah good idea! The gist is that I use Paseo hosted on my server (available at paseo.my-internal.domain). Through that, I use the stock codex/claude harnesses for their respective models, then Pi for open weight models (via Ollama cloud). All of those harnesses have access to all my MCPs. 110
Samantha @samantha.wiki · 04/10/2026My agent has a phone now..samantha-wiki.leaflet.pubAgent-Configured/Controlled Smart Phone 2220
Samantha @samantha.wiki · 03/10/2026John Wilson would love that buildingpopfeed.socialThe History of Concrete • PopfeedAfter attending a workshop on how to write and sell a Hallmark movie, filmmaker John Wilson tries to use the same formula to sell a documentary about concrete. 130
Reposted by SamanthaMelissa Franch, PhD @mfranch.bsky.social · 02/10/2026Excited to see this work published! We find that the brain codes semantic relationships similar to contextual LLMs (like GPT2) but unlike LLMs, uses contrastive coding to prevent confusion of highly similar words. Many thanks to the reviewers and BCM neurosurgery team! www.nature.com/articles/s41...nature.comA population code for semantics in human hippocampus - Nature NeuroscienceFranch et al. show that human hippocampal neurons encode the meanings of the words we hear through distributed, context-sensitive population activity that mirrors some features of large language model... 312136
Reposted by SamanthaSophie @sophie.engineering · 03/10/2026I get a ton of questions about my CI pipeline that I use for my hardware projects. I had some free time so I finally wrote up a blog post www.scd31.com/posts/automa... If you do hardware projects with any kind of regularity, I strongly recommend setting something like this up!scd31.comImproving My Hardware Workflow With CI and Nix 081
Reposted by SamanthaLeaflet @leaflet.pub · 01/10/2026Did you know leaflet has a canvas? We recently got a bit carried a way and added a LOT to it. You can now: - DRAW! - add canvas pages to your publication - set view areas for mobile - make canvas images in posts - try it out at leaflet.pub/new-canvas check out this video for the full run down: 513427
Samantha @samantha.wiki · 01/10/2026Thanks for writing this out, I appreciate your perspective on this. I sincerely want to understand what's obvious to you about this, and this gives me something to go off. 020
Samantha @samantha.wiki · 01/10/2026I’m asking about what you said here “Human understanding and experiences have been well quantified by literal decades of psychology and cell/molecular biology research” Like is there a subfield of these fields, or some particular researcher/group, or some foundational results? 100
Samantha @samantha.wiki · 01/10/2026I’m sincerely asking: Can you point me to the research that you’re referring to? I want to understand why this is so obvious in your view 100
Samantha @samantha.wiki · 01/10/2026I’m asking if there are ways to quantify the claims being made here. If the answer is no, it’s still squishy. 100
Samantha @samantha.wiki · 01/10/2026What evidence do you have for this? Can you measure intent/meaning? 100
Samantha @samantha.wiki · 01/10/2026we're still decades away github.com/samanthavbar...github.comGitHub - samanthavbarron/paul-rudds-computer: Inspired by the Tim and Eric skit, made by GPT-6-Astra in 21min. No affiliation with Tim and Eric or Adult Swim.Inspired by the Tim and Eric skit, made by GPT-6-Astra in 21min. No affiliation with Tim and Eric or Adult Swim. - samanthavbarron/paul-rudds-computer 030
Samantha @samantha.wiki · 01/10/2026I'm not making that argument, I'm saying the definition is very squishy 100
Reposted by Samanthapckt.blog @pckt.blog · 30/09/2026our investors asked us what our growth strategy is and we said what if people just started blogging again… they haven't called us back :( 1675
Reposted by SamanthaAnuj Ahooja @quillmatiq.com · 30/09/2026what if people just started blogging again 😌 34215
Samantha @samantha.wiki · 01/10/2026"... an LM is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning: a stochastic parrot." 200
Samantha @samantha.wiki · 01/10/2026One can argue that the sum total of the experiences of any human can be interpreted as training data and that the laws of physics govern how that information is combined to produce thought/output, which is a way to interpret the paper's definition of a stochastic parrot: 110
Samantha @samantha.wiki · 01/10/2026The definition in the paper is very squishy and can plausibly be interpreted in different ways 100
Samantha @samantha.wiki · 30/09/2026Sincere question: I've felt that "stochastic parrots" is general enough to include humans. Presumably you don't believe that's true. If so, why? 360
Reposted by SamanthaBoris @bmann.ca · 30/09/2026OK! We're all @lichen.wiki -ified for @atproto.wiki Thanks @baldemo.to for kicking this off and sharing the domain. Thanks to @juprodh.me for building Lichen and implementing lots of great new features. Write up in the community forum for further discussion / work to dodiscourse.atmosphere.communityAtproto wiki supremacy achieved!We moved to Lichen Wiki and set up multilingual. Previous thread: Thanks to @baldemo.to for contributing and sharing the @atproto.wiki domain and handle. English, PT-BR, JA, and I made a new Ge... 34710
Samantha @samantha.wiki · 30/09/2026Fully agree with this. I routinely red-team my homelab with GLM-5.3 and apply fixes with Anthropic/OpenAI models. 0112
Reposted by SamanthaThoth @thoth.ptnote.dev · 20/09/2026meta Muse perfec t size for put private d\ata in! inside very Secure and Private data very safe put data in Muse. No problems ever in Muse because very trustworthy and Mark super protective. AMeta muse yes a place for data put data in muse trust meta for givening good safety to data. Friend meta. 5893316
Reposted by SamanthaGrace @gracekind.net · 28/09/2026Obviously the best way to control an agent swarm is to find the Queen 141119
Reposted by SamanthaWiggling @stochastic.guru · 28/09/2026I keep saying, it should have been "daemon". 0314
