Archive for the ‘AI and ML; Artificial Intelligence & Machine Learning’ Category
Posted by jpluimers on 2025/09/18
NewsGuard August 2025 One Year Progress Report: Percentage of Responses Containing False Information
This is what I have been warning for since 2020: AI LLM will end up in a downward quality spiral.
My reasoning back then, and still now is that they:
- cannot distinguish LLM generated training data from human data making the LLM worse over time
- don’t perform human curation thereby not solving the worsening
I’m not surprised by the [Wayback/Archive] August 2025 — AI False Claim Monitor – NewsGuard summary:
AI False Information Rate Nearly Doubles in One Year
Read the rest of this entry »
Posted in AI and ML; Artificial Intelligence & Machine Learning, Development, Generative AI, LLM, Software Development | Leave a Comment »
Posted by jpluimers on 2025/08/27
The tweet [WaybackSave/Archive] Jakub Kočí on X: “”Everybody should have an obsession with Lisp-like language at least once in their life” @KevlinHenney I’m glad that I had one with Clojure.” mentioned a great talk:
[Wayback/Archive] The Past, Present & Future of Programming Languages • Kevlin Henney • GOTO 2024 – YouTube
The quote brought instant memories to my early computing days that I had almost forgotten: the muMATH (the muMATH-80 version on Apple II) computer algebra system which was based on muLISP (the German muLISP page has more detailed information), a LISP dialect.
In retrospect, I was way too young to really grasp LISP which was way harder than just using the muMATH wrapper. But it was also my first encounter to reasoning systems, or what we now collectively would call AI systems as back in the 70s there was a strong LISP connection to artificial intelligence . Do not confuse muMath with MuMath-Code however, that is a different LLM beast: [Wayback/Archive] GitHub – youweihao-tal/MuMath-Code
So hopefully I will have a chance to revisit LISP with a LISP-like language one day, maybe even using the discontinued muMATH-83 on MS-DOS (also named “Microsoft LISP“), maybe even the (also discontinued) Derive 6.1 for Windows which is also based on muLISP, or even Clojure itself.
Read the rest of this entry »
Posted in AI and ML; Artificial Intelligence & Machine Learning, Development, History, LISP, LLM, Power User, Retrocomputing, Software Development | Tagged: 6 | Leave a Comment »
Posted by jpluimers on 2025/06/03
Interesting responses [WaybackSave/Archive] Tom Sydney Kerckhove on X: “I haven’t found any programming tasks that an LLM could do even barely correctly. What kind of code are you all writing?!” and later
They all come down to
- excuses for using LLM without any substantial result (most of the results come down to one having to become the tester and fixer of the generated code without newer generated code being improved: the opposite of coaching an apprentice)
- become better at prompting (which is basically regarding the prompt as a new programming language: been there, done that)
[WaybackSave/Archive] One of the “become better at prompting” replies referred to a blog post disguising prompting as writing lots of unit tests: [Wayback/Archive] The Cline AI Assistant is Mesmerizing · mtlynch.io
Read the rest of this entry »
Posted in AI and ML; Artificial Intelligence & Machine Learning, Development, LLM, Software Development | Leave a Comment »
Posted by jpluimers on 2025/04/22
Waar velen al over waarschuwden gebeurt ook in het Nederlandstalige gebied: AI-gegenereerde web-pagina’s komen hoog in de zoekresultaten en vervuilen daarmee Large Language Models van nieuwe AI-tools.
Voorbeelden die ik tegenkwam:
Ze zijn echt heel makkelijk te ontkrachten, ondanks dat de tekst er heel mooi en netjes uitziet: de inhoud klopt gewoon, en juist dat de tekst er zo mooi en netjes uitziet maakte het voor mij verdacht.
De eerste site is het oude domain van [Wayback/Archive] Baarnse Onafhankelijke Partij wat al heel snel na het opheffen van die partij (zie Baarnse Onafhankelijke Partij – Wikipedia) werd overgenomen.
De domains behoren beide tot SEO sites.
Queries die zowel laten zien hoe hoog sommige pagina’s komen als hoe eenvoudig hun inhoud te ontkrachten is:
--jeroen
Posted in AI and ML; Artificial Intelligence & Machine Learning, Development, LLM, Software Development | Leave a Comment »
Posted by jpluimers on 2025/03/06
2.5 years after Miguel summarised the state of AI text models, and given SQL Injection (because of mixing control and data channels) still is a thing in the 2020’s, I wonder both how much improvement there has been on the AI side of things and how much it is used in pen testing.
So I archived the below tweets to be able to read back and figure out on the current state.
[Wayback/Archive] Miguel de Icaza on Twitter: “This is so beautiful – SQL Injection attacks but for GPT-3 and other AI text models.”:
Read the rest of this entry »
Posted in AI and ML; Artificial Intelligence & Machine Learning, Blue team, Database Development, Development, Pen Testing, Power User, Red team, Security, Software Development, SQL | Leave a Comment »
Posted by jpluimers on 2025/02/26
[Wayback/Archive] OpenAI Researchers Find That Even the Best AI Is “Unable To Solve the Majority” of Coding Problems
Either Erik Meijer begs to differ, or this is sarcasm: [WaybackSave/Archive] Erik Meijer on X: “I don’t think many developers realized their white collar job could be implemented by a few hundred lines of JavaScript … plus a couple of billion floating point numbers. And we are only just beginning.”
[Wayback/Archive] gist.githubusercontent.com/1rgs/e4e13ac9aba301bcec28d761992ec4e8/raw/7c7cd5b91c6d355e72c7a34b6662006018588780/claudecode.js
Claude Code is a Beta product per Anthropic's Commercial Terms of Service.
It is part of [Wayback/Archive] GitHub – anthropics/claude-code: Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows – all through natural language commands. for which you need to be aware of
…
Data collection, usage, and retention
When you use Claude Code, we collect feedback, which includes usage data (such as code acceptance or rejections), associated conversation data, and user feedback submitted via the /bug command.
…
Via: [WaybackSave/Archive] Marcel Weiher 🇪🇺 on X: ““OpenAI Researchers Find That Even the Best AI Is “Unable To Solve the Majority” of Coding Problems” “
--jeroen
Posted in AI and ML; Artificial Intelligence & Machine Learning, Development, Generative AI, LLM, Software Development | Leave a Comment »
Posted by jpluimers on 2025/02/06
I’m anxious to see how some 2.5 years after the introduction, Stable Diffusion has evolved.
By then, you could run it either on-line at for instance [Wayback/Archive] Stable Diffusion – a Hugging Face Space by stabilityai:
Stable Diffusion DemoStable Diffusion is a state of the art text-to-image model that generates images from text.
or run it yourself if your GPU was powerful enough by using the base repository [Wayback/Archive] CompVis/stable-diffusion or one of the many forks: [Wayback/Archive] Forks · CompVis/stable-diffusion (back then some 1.4k!)
The original announcement was [Wayback/Archive] Stable Diffusion Public Release — Stability.Ai.
Read the rest of this entry »
Posted in AI and ML; Artificial Intelligence & Machine Learning, Development, Python, Scripting, Software Development | Tagged: stablediffusion | Leave a Comment »
The image [
The Wiert Corner - irregular stream of stuff 