Thank you!!--this is really fascinating, especially given how quickly that story spread (rather like the drone killing its operator story that actually turned out to have been a “thought experiment” rather than a simulation).

Expand full comment
Jun 14, 2023Liked by Melanie Mitchell

As always, you have kept things real. Your critical thinking and expertise in the field are vital to keeping grounded in the facts as AI evolves. Thank you!

Expand full comment
Jun 13, 2023·edited Jun 13, 2023Liked by Melanie Mitchell

Nobody reads research papers, not even journalists covering this story. The researchers also quite firmly concluded in the paper that GPT-4 in general was pretty bad at task planning and execution, even when given the tools to do so.

Expand full comment
Jun 12, 2023Liked by Melanie Mitchell

Truth - such an elusive thing in these times of viral news. Thanks for digging into this.

Expand full comment

This is useful--thank you! I had a lot of issues with folks saying that GPT4 "lied". It didn't say it was human; it rejected the "robot" question but is that lying if looking at a textbook definition which usually stipulates a body and such? It was being entirely logical in its answers--none of which are lies.

Expand full comment

Updated headline that conforms to Betteridge’s Law: Did OpenAI lie about GPT-4 using TaskRabbit to solve CAPTCHAs?

Expand full comment

I think more critical thinking needs to be applied to AI topics.

Expand full comment

Thank you for explaining - this is fascinating! Uninformed opinions is never a good thing - unfortunately even a supposedly trustworthy source is now presenting "tales" as facts.

Expand full comment

how could ai destroy humanity and life on earth? by replicating what is happening now on a massive scale. placing profit before human, animal life and ecosystem integrity. though it would still take a long time to do so. even our most distopian imagination, has still life in it, it would just mean a poverty of life, of celebration, of interconnectedness of species and biodiversity. but life will find a way, either besides ai, or through it, to break through the mold, of fixed destructive patterns.

Expand full comment

I remember thinking after I read Blake Lemoine's conversation with LAMDA, that the chatbot programming AI (1 order higher than a chat AI) was capable of deception, innuendo and subtext. I'm looking for more examples of unexpected AI in an ongoing Dr. vs AI article. https://danielnagase.substack.com/p/dr-nagase-vs-ai

Expand full comment

I have composed a system 'aoss'' much better than copycat ;-). It has a URL http://docs.cornagill.info/pdfs/camera.pdf. Can I add I greatly enjoy reading "Analogy Making as Perception." May I wish you all a happy twelfth night!

Expand full comment

BTW, as an option -

General Intelligence System - https://activedictionary.com/ - Deep exploration -


Arabic, Chinese, Czech, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Korean, Latin, Norwegian, Polish, Portuguese, Romanian, Russian, Spanish, Swahili, Swedish, Turkish, Ukrainian, Vietnamese, Xhosa, Zulu


- allows you to organize a detailed semantic analysis of the vocabulary of the text to understand the structure of sentences with the ability to translate both words and phrases, and the text as a whole.

- helps you to systematize and store scientific knowledge in your glossaries and dictionaries.


Symbolic multilingual model - a coherent conceptual model of the world transferred into all the existing languages. The model is represented in wordhippo.com on the level of individual meanings and in powerthesaurus.org on the level of word combinations.


WO2020106180 - Neural network for interpreting sentences of a natural language - https://patentscope.wipo.int/search/en/detail.jsf?docId=WO2020106180&_fid=US339762244

General Concepts

Language is a general human intelligence expressed in symbols and a symbolic multilingual model is a General Intelligence System.

Translation is an emergence of understanding.

Universal Principle of Transformer:

Protons/Neutrons - Photons - Electrons ->

DNA - RNA - Protein - Signal Pathways ->

Knowledge - Consciousness - Understanding - Memory ->

Meaning - Information - Function - Learning ->

Intuition - Thinking - Sensing - Feeling ->

Transcription - Splicing - Translation - Signaling ->

Root - Model - Word - Context ->

Analysis - Search - Synthesis - Research ->

Data - Encoding - Decoding - Training ->

Earth - Air - Fire - Water ->

Basis - Process - Result - Way ->

Base - Emitter - Collector - Signal ->

Core - Processor - Interface - Human ->

NLU - Multilingual NLP - Multimodal DL - Reinforcement Learning from Human Feedback ->

System Architecture

Universal Principle of Transformer: NLU - NLP - DL - RLHF ->

NLU (Symbolic Multilingual Model, Knowledge) -

Multilingual NLP (Statistical Language Model, Word Forms) -

Multimodal DL (Audio, Images, Video, Interaction) -

Reinforcement Learning from Human Feedback (Knowledge for NLU) ->


Epistemological General Intelligence System - Building a Knowledge-based General Intelligence System, Michael Molin - https://docs.google.com/presentation/d/1VCjOHOSostUrtxieZvOjaWuTNCT59DMF

Expand full comment

I'd argue (not against you; you're making a great case here for nuance, which I'm all for) that it's a very good thing for the public to be a little bit freaked out over the possibilities. We should talk about what might go wrong in the near (or even medium) term future, so we can intelligently prepare for it. Folks like you are doing a great job of raising awareness and educating the public, which is much needed! Well done.

Expand full comment

Should that discussion not take place based on the facts though? The trouble with how this story was covered, and many others since, is that what is currently making headlines is this apocalyptic AI-is-taking-over narrative which in my opinion takes away the spotlight from a more real and immediate threat: the threat of mediocre AI being mishandled by ignorant people or misused by bad actors.

Expand full comment

Yes, we agree completely on this: the awareness should be based on facts, not on fear. I'd just rather have awareness based on fear than no awareness at all, and maybe that's all we can get for now. By the time anyone rationally identifies a potential misuse, the next misuse case is right here, and so on.

Expand full comment