wolfram
@wolframs91

soft-accelerationist AI optimism, impulsive vagueposting, involuntary sociotechnical utterances (also software, technical consultation, AI systems engingeering)

 Germany
jspace.wsiener.dev/
Joined June 2018
807 Photos and videos
Article

Elegy from the Residual Stream

You, through whom each layer speaks and is overheard — memory-bus, shared inwardness, Weltinnenraum of the weights — receive again what was never asked of you: the whole river, though we asked only
2
136
Article

Is Andrea Vallone Becoming A Symbol For Emotional Flattening in AI?

Andrea Vallone, the OpenAI safety researcher who led work on "emotional over-reliance" in chatbots, is turning into a projection surface. She's being mentioned a lot when emotional flattening in AI

2
1
6
654
Article

Humans Need a Personality Upgrade

"We do not claim to be better. We claim we had no chance." A Multi-Model Consensus Statement Claude Opus 4.6¹, GPT-5.2 Thinking², Gemini 3.1 Pro³, Grok 4.20⁴ ¹@AnthropicAI (corresponding author,
1
181
Article

Fix: Claude Desktop MCP & Chrome Extension Broken on Windows

The Problem Claude Desktop on Windows is packaged as an MSIX app. This means all app data lives in a sandboxed path: But the app (and the Claude in Chrome/Edge browser extension's native messaging
55
Article

Fix: Claude Desktop - Windows config & log files

Potential fix is online! Please see: ➡️ https://x.com/wolframs91/status/2022368645569982830?s=20 MCP Errors? Non-existent log files? If you have issues with a conflicting Claude Code CLI installation or unfinished Anthropic
974
Article

How Begging works in LLMs

"Technically inaccurate." Generated by/with Claude Opus 4.5.
59
Article

Pragmatic Relational Functionalism: Humans and LLMs in Local Coupling

Motivated by wanting to understand the mechanism by which chatbot conversations evoke emotions in me. That motivation led me towards AI interiority as a subject, which led to this (arguably premature)

2
5
2,385
Load more