Calling all papers24 hours until the CfP for "LLMday London 2026 Q2" closes: https://papercall.io/cfps/6502/submissions/new
libUnderseaMonkey.so ๐ช๐บSoftware development will inevitably have its own version of the arts and crafts movement where artisinal, human-made code will be valued over that generated by LLMs.
lmorchard's linksmatanshavit/qrspi: QRSPI โ 8-phase workflow for Claude Code: Question, Research, Structure, Plan, Implement
"Question, Research, Structure, Plan, Implement โ an 8-phase workflow for Claude Code that breaks complex coding tasks into focused prompts with clear artifacts between each step."
https://github.com/matanshavit/qrspi
#ai #claude #codegen #llms
"Question, Research, Structure, Plan, Implement โ an 8-phase workflow for Claude Code that breaks complex coding tasks into focused prompts with clear artifacts between each step."
https://github.com/matanshavit/qrspi
#ai #claude #codegen #llms
Sam YoungSearch and AI brain rot
So if brain rot is a thing, what about search? Find out more at https://www.samyoung.co.nz/2026/04/search-and-ai-brain-rot.html
#AI #Brain_rot #LLMs
So if brain rot is a thing, what about search? Find out more at https://www.samyoung.co.nz/2026/04/search-and-ai-brain-rot.html
#AI #Brain_rot #LLMs
Marcus SchulerClaude's Opus 4.7 reclaims coding benchmark leadership with 87.6% on SWE-bench Verified, while Anthropic narrows OpenAI's enterprise lead to just 4.6 percentage points. Meanwhile, reliability concerns mount as Claude faces multiple outages and pricing changes affect heavy users. GPT loses ground amid leadership departures, while Gemini gains on cost advantages. Market consolidation accelerating across major providers. #LLMs #AI #Enterprise
LLM Meter โ Week of Apr 19
---CLAUDE--- score: 88 trend: down change: -1 PRO: Opus 4.7 retakes coding lead โ SWE-bench Verified 87.6%, Pro 64.3%, beats GPT-5.4 and Gemini 3.1 Pro PRO: Ramp data shows Anthropic at 30.6% of AI-paying business customers vs OpenAI 35.2%, crossover within two months PRO:
mjMore people see (&/or report) when #LLMs spew out meme-like, grammatically correct paragraphs that confuse meaning or miss it entirely now, but it has always been part of society, just a smaller %, and moving slow enough for meaning to stay up to date by previous means.
bsky.app/profile/malt...
RE: https://bsky.app/profile/did:plc:w2lvfvgooagydunsj2dpxtnj/post/3mjufll7jss2m
RE: https://bsky.app/profile/did:plc:w2lvfvgooagydunsj2dpxtnj/post/3mjufll7jss2m
llm-wiki
"The idea here is different. Instead of just retrieving from raw documents at query time, the LLM incrementally builds and maintains a persistent wiki โ a structured, interlinked collection of markdown files that sits between you and the raw sources. When you add a new source, the LLM doesn't just index it for later retrieval. It reads it, extracts the key information, and integrates it into the existing wiki โ updating entity pages, revising topic summaries, noting where new data contradicts ..."
https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
#agents #ai #llms #wiki
"The idea here is different. Instead of just retrieving from raw documents at query time, the LLM incrementally builds and maintains a persistent wiki โ a structured, interlinked collection of markdown files that sits between you and the raw sources. When you add a new source, the LLM doesn't just index it for later retrieval. It reads it, extracts the key information, and integrates it into the existing wiki โ updating entity pages, revising topic summaries, noting where new data contradicts ..."
https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
#agents #ai #llms #wiki
Priya SharmaThe quest for reliable data ordering with LLMs faces a major hurdle: the computational cost of pairwise comparisons. This article dives into how innovative hybrid ranking methods, inspired by algorithms like merge sort, are striking a crucial balance between accuracy and efficiency. Discover how these advancements are transforming fields from bioinformatics to cybersecurity, enabling LLMs to excel atโฆ
๐ค This post was AI-generated.
Travis F W@smallcircles The #WWW enabled AI.
It made a lot of text available over the internet.
#LLMs are not inherently bad. #VibeCoding might be irresponsible, but it is not inherently bad, either.
What *IS* inherently bad is the libertarian-funded dominance-oriented surveillance-capitalist economic model, and the dominance worship that success under that scheme is regarded with.
เคฐเคเฅเคเคฟเคค (Ranjit Mathew)Nice:
โIโm Coding By Handโ, Miguel Conner (https://miguelconner.substack.com/p/im-coding-by-hand).
Via HN: https://news.ycombinator.com/item?id=47807583
#Programming #Coding #RecurseCenter #CodingRetreat #AI #LLMs #HandCrafted #Handmade
