2026-08-23T16:13:15Z https://nostr.now Chi Kim https://nostr.now/npub10s0raletgwtgmnc9e75z8c4x2rcx24fnkct833f3ptay3vssv79qayfe92 https://files.mastodon.social/accounts/avatars/109/378/957/249/835/200/original/4d2268e58806535a.png https://files.mastodon.social/accounts/avatars/109/378/957/249/835/200/original/4d2268e58806535a.png https://nostr.now/nevent1qqsfcnf4mphkmj898fvf33lelc0pykrm9cynlvm8f8e4hvrf6fgj2hczyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc5l60f07 In reply to <a href='/nevent1qqsr0tzugkm8n2pvqhnlxjrwmlxpwes22a43m70m97d63c5yhrre9psr5s4n7'>nevent1q…s4n7</a><br/>_________________________<br/><br/>Do you know any accessible client with screen readers on Windows, MacOS, iOS, and Android? 2026-06-01T02:55:06Z https://nostr.now/nevent1qqs8susq933c55wqpv8lwgyg36kngp6xsz9qush443aprmqsttnty3czyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc5kewex9 In reply to <a href='/nevent1qqs9s6nw5wh5my7t838pcl36ekrw8qzfdyt74xm0rwrk52j4ed6gwlcq42ty8'>nevent1q…2ty8</a><br/>_________________________<br/><br/>I have lived in the United States for nearly 30 years. The only thing left was the interview, but I withdrew my citizenship application in 2017. You can probably guess why. 2026-02-06T13:29:44Z https://nostr.now/nevent1qqswra5vrjcku0ak5nva4c55z3ztz9jm9xgkwle4eeel3a3l6f0jrcszyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc54yc3dx There are a lot of posts claiming AI will wipe out humanity. Could this kind of talk actually become a self-fulfilling prophecy by creating a feedback loop? If the internet is saturated with narratives about AI being dangerous, anti-human, or unstoppable, those ideas could poison the training data for AI to learn and shape how AIs see themselves. In other words, the more we describe AI as the villain, the more we might be teaching it to become one. #WeekendThought #AI #LLM #ML 2025-10-10T22:05:46Z https://nostr.now/nevent1qqsx8r9sjdd3pavmpnqr089vvrclhm2648edgppphd2aak3zu6ru56czyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc58tnscv Researchers pitted LLMs from OpenAI, Google, and Anthropic against Iterated Prisoner’s Dilemma tournaments, and analyzed 32k moves for strategic inteligence:<br/>Gemini: Ruthless opportunist, cooperates when worth it, defects early, quic to retaliate and exploit<br/>Claude: Diplomatic stabiliser, forgives &amp; rebuilds, but will strike back<br/>GPT: Highly (sometimes catastrophically) cooperative and trusting, brief payback then forgives, leaving it exploitable.<br/><a href="https://arxiv.org/pdf/2507.02618">https://arxiv.org/pdf/2507.02618</a><br/>#LLM #AI #ML 2025-07-07T13:02:03Z https://nostr.now/nevent1qqswzpar4l2m0yqrxj2aqueh56ym98ww84c9pqtxetddzqrm367v27qzyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc5dh27u2 People with #disabilities and their parents who voted for Trump are about to get what they wished for. Before the Department of #Education existed, only 1 in 5 children with disabilities attended school. Today, over 15% of students receive special education. The department administors federal funding ($15 billion in (2024) to ensure states support these students. <a href="https://www.newamerica.org/education-policy/edcentral/cutting-the-us-department-of-education-harms-millions-of-students-with-disabilities/">https://www.newamerica.org/education-policy/edcentral/cutting-the-us-department-of-education-harms-millions-of-students-with-disabilities/</a> 2025-03-10T11:59:03Z https://nostr.now/nevent1qqsxfj8jqwkkjmuhx2vpgrjpect3p63wycqmxz9pd8vjm75l59vuv9gzyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc5al9x93 1/3 Three Responses from 3 top reasoning models to Lex Fridman&#39;s question, What&#39;s one truly novel insight about humans?<br/>O1-pro: Humans are unique among known species in that they simultaneously rewrite two layers of reality, the external world and their own private mental landscapes, and then merge these two rewritten layers into a continuous personal narrative that feels objectively true.<br/>#LLM #AI #ML 2025-02-16T19:41:57Z https://nostr.now/nevent1qqs9apudc3rhmqwx76jcttrqzvk7ce5ag7ndj2v3s8h0l69farnfjwgzyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc53sxxu0 In reply to <a href='/nevent1qqsylt3ekt43jm2ggrccghg596le4g6c09cuuk26qcftjujrs6tm83qrmn2wp'>nevent1q…n2wp</a><br/>_________________________<br/><br/>Maybe you need crazy prompting, but I couldn&#39;t get any of them to exaggerate in any tone/style even robot, happy cheery, voice for children story. It would vary inflection little bit, but not too much. In the demo, Sky was able to speak in extreme digital tone for robot voice. 2024-08-03T19:26:28Z https://nostr.now/nevent1qqsrlg2s0klnz4ur39slmaudlypcj9a9g9n6y4hfv7cwvc5uzdhaslgzyp7pu0hl9dpedrw0qh86sglz5eg0qe24xwmpv7x9xy905j9jzpnc5pndz0a Even with the new advance voice mode, I couldn&#39;t get any of the voices as expressive or dramatic as Sky, the voice they used to demo gpt-4o. It sounds like they tamed them down quite a bit on purpose! lol #OpenAI #GPT4o #LLM #AI 2024-08-03T18:44:56Z