LLM
concept · 29 episodes · JSON
Also known as: LLMs
- Just the Links … I wrote that! Not an LLM! Ugh, I feel gross. FIRST. Ars Technica: A missing underscore sent innocent man to prison for 18 months 1 is a story …
- The Irritating Aspect of Spiciness; Improving Pharmacies Somewhat … It is weird to people because computers are supposed to be programmable and “smart” but the thing to understand here is that when you ask an LLM to …
- For All Humanity?; Gordon Moore Folds Your Clothes … Either LLM model weights aren't copyrightable, in which case distilling an LLM cannot be copyright infringement; or they are the product of mass copyright infringement themselves, in …
- A Different Strategy, The Sanest Choice In An Insane World … In an insane world, it was the sanest choice. (Unlike contemporary LLMs, there’s not that much evidence that Arnold the Terminator would encourage John to die by …
- Anything LLMs Can Do, I Can Do Better … I won’t say LLMs can’t do that because one should never say never, and I also won’t say LLMs can’t do that because anything …
- That Whole Moltbot/OpenClaw YOLOing AI Agents Thing … But it’s because the LLM isn’t actually running a program (which can be good at counting!), it is, well, being an LLM about it. This is …
- The Problem is Defining the Problem; The Unreasonable Effectiveness of Caring … And that's quite the opposite of what LLM assistants promise. 3 I think the point here is that agentic coding/LLM-assisted coding (and therefore I think …
- Things Could Have Been Different; LLMs as LitTech, Their Use as Response to Mass Functional Illiteracy … 1.1 LLMs as LitTech, Their Use as Response to Mass Functional Illiteracy I will start with some of the interesting bits first to get your attention before …
- An End Of Year Opinion About AI Because Why Not; Good Enough Mitigation of Reasonably Foreseeable Harm … So I’ll just be quick and pretend that we’ve covered that “AI” can mean lots of things, “generative AI” is something specific, people are using LLMs …
- Efficiency Gains; I Understand You, But Could Never Be You; Can You Teach A Machine To Be Human? … 1.0 Some Things That Caught My Attention 1.1 Efficiency Gains It is a bit of a cheap shot to say that LLMs aren’t great. That …
- Not About How it Looks … Maybe Grok (or whichever LLM) is less biased, or knows about more experts in whichever field? My gut feel is that Grok might “know” more experts, but that …
- se20e05: AI Models; The All Souls Exam; Melvynn Bragg as an Oxford Don; Grounding; Being Human; Worry About The Same Things … 1.0 Some Things That Caught My Attention 1.1 Should Anonymous Posting Online Be Forbidden? I wrote back in s20e01 1 about how current LLMs/token predictors …
- Better than Average; How People Work … posted an observation a few weeks back that went a bit like this: LLMs / generative AI / token prediction machines are trained on a mass of data They are …
- Your Model is Naive and it’s Not The Territory … The models are complex enough when encoded in legislation and then a business rules engine, but now the models are potentially implicitly encoded into your LLM by you …
- Do Reply; Use plain language, and tell the truth … The reason why LLM text for me is bad is that it’s insipid, which is not a plain language word to use, but the secret is to …
- Do Not Reply Stickers; Computers were a Bad Idea; Two Things About Startups; and more tidbits … voice-based LLM interfaces. (Of which! Voice-based LLM interfaces! Such a hot mess right now! A lot of the problems in voice interfaces like lack of discoverability …
- What Are We Doing Here, Exactly?; LLMs Are Playful, Actually … 1.2 LLMs are Playful, Actually Oh jeez where to begin with this one-slash-what a doozy. Two things coming together in my mind here. Here’s …
- Justice in Forensic Algorithms; The One With Retroactive Bereavement Fare Claims; Bits and Pieces … Here’s the good bits, which is mainly me doing a précis (sorry, organic, sentient LLM summarization) of the paras 24-32 of the judgment: The judge helpfully …
- Context is King; Explicit People Personality Preferences; OS Licensing, but Different … copy to be! I guess people will create the different versions with LLMs. (BAD LLM! NOW DO IT 10% LESS PUMPINGLY) Now I’m sad. 1.2 OS …
- What does it take to trust?; The Young Lady’s Illustrated Tour Guide … That time it was about how the Rabbit R1 hardware-enclosed LLM assistant had struck peoples’ attention, and whether the demoware showed actually-useful, real-world examples. Anyway …
- It’s not a hardware or software problem, dummy … Rabbit and LLMs are a kind of symptom of this because they show the promise -- whether realized or not -- of being able to take tedious, siloed tasks that …
- How old things keep living; Impossible AI; Look, just call it AI … Case Studies | Artificial Intelligence | JAMA Pediatrics | JAMA Network ( archive.is ) ↩︎ What I Learned Using Private LLMs to Write an Undergraduate History Essay – zwischenzugs ( archive.is ), 27 December, 2023 ↩︎
- Not Up and To The Right; Just Deeper, Really; Better Wording Faster; Tools for Remembering; A Centaur for Caring?; Of Course They're Art … Flynn, continuing my conversation about how LLMs and other so-called artificial intelligences can be used by people. There was this concept of caring and relationship management. When …
- The Torment Nexus Section; RFPs, man. RFPs. … OK, so Wired did a really interesting / horrifying story about how LLMs (automatic text generation engines and word/phrase-relationship engines) have upended the whole business of responding …
- Tech Journalism, But Not Tech Journalism; Sounds More Like A You Problem … Just the other week, The Atlantic published a tool for searching the Books3 dataset [4] used to train generative LLMs, although the not-so-great news is that …
- Marks on Flat Planes … Fine-tune your LLM character-based model on all your Sherlock Holmes data so that when you see your dailies of your Sherlock Holmes piece of content , you …
- Someone Comes to Town, Someone Leaves Town; Streaks Broken; Aigg on your Face, Big Disgrace … One reason why, I'm pretty sure, is that LLM-generated text is super easy to glaze over entirely because it's so predictable and we're already …
- To help you best, I'd need to know you … Henley says the potential of LLMs is that they could feed the relevant context to the model behind the scenes and use that to preemptively suggest what I …
- This was not the future we were promised^w sold … Though engineer seems too formal, it immediately conveys the point I’m making: “Prompt engineers use natural language prompts and LLMs to produce new content” @AndrewKemendo Caught my …