It was an okay post. I think I wasn't careful enough with some nuances. Anyway, reading some of the response I realized that both topics (AGI and self-driving) are so charged and toxic right now and I should just stick to technical things
SAFETY
-
GPT-4 limitations for large-scale data automation use cases
By
–
I have no objections to using automation to do large-scale data processing. But it needs to be developed & tested for the use-case. "zeroshot" (ugh) or not, GPT-4 isn't that. You don't even know what its training data is.
-
Deepfake News: Which Politician Is Most Vulnerable?
By
–
“Deepfake” news? I think there may be a politician that is more likely than others.
-
Stanford Team Researches Foundation Models Applications and Social Impact
By
–
Looking for more research on foundation models? This team is studying the applications, technical advances, and social impact of these powerful models – from improving state-of-the-art methods to evaluating the models' downstream uses: https://
stanford.io/48Uc6BI -
Mitigating LLM Hallucinations: RAG and Prompt Tuning Strategies
By
–
What are some mitigation techniques against LLM hallucinations? We have found that RAG combined with some prompt tuning usually works best. By giving the LLM strict instructions, it will respond with content that is sent in the context. At Abacus AI, all our deployed LLMs in
-
AI for Defense and Intelligence: Book Review by Pat Biltgen
By
–
I just finished "AI for Defense and Intelligence" by Pat Biltgen. Pat is widely known in intelligence community circles, he is a frequent speaker at AFCEA Intelligence Community and INSA events and a frequent poster on LinkedIn. His book does a fantastic job reviewing what is
-
Science Strengthens Image Integrity Controls for Published Research
By
–
[#Article] How the Science journal ensures image integrity in its publications https://actuia.com/actualite/comment-la-revue-science-veut-sassurer-de-lintegrite-des-images-dans-ses-publications/
… #AI #artificialintelligence -
ChatGPT fails to read child’s handwritten chess moves accurately
By
–
Tried ChatGPT to read my 7-year-old’s handwritten chess move sheet and it failed miserably to correctly read all moves except the first one. But sure, AGI is just around the corner.
-
Patchscopes: Framework for Explaining LLM Internal Representations
By
–
8/ Patchscopes – proposes a framework that leverages a model itself to explain its internal representations; it can be used to answer questions about an LLM’s computation and can even be used to fix latent multi-hop reasoning errors.
-
AI Self-Improvement Breakthrough: Path to AGI Emergence
By
–
"AI may become the first technology with the means to improve on itself." Hard to think of a more consequential moment than this.