Steve Chambers

Comment Threads

• Martin Rodgers commented on Made of lies (and more lies)
  Again, I'm talking about search results. I've made that point well enough already. I see no point in repeating myself....
• Martin Rodgers commented on Made of lies (and more lies)
  One way or another, the next step seems to reading more history....
• SFReader commented on Made of lies (and more lies)
  Re: '... bellingcat's findings' Not familiar with them so looked them up on Wikipedia - seems like a legit news org. Also checked a couple of sources/citations on that Wikipedia page - Bellingcat won the European Press Prize in 2019. https://www.europeanpressprize.com/shortlists/year-2019-2/ FYI - Years ago Wikipedia was tested vs. Britannica for science info: Britannica was only slightly better -- the testers found 4 errors on Wikipedia. Geeze - who knew that open-sourcing/citizen science writing could be such an effective means of gathering and distributing info/knowledge! My guess is that Wikipedia also has a bunch of reasonably good editors who sort...
• DeMarquis commented on Made of lies (and more lies)
  You made an active claim (that Bellingcat is unreliable). I am now asking you to back that up. If that's trolling then I couldn't be more proud to be a troll. And, no, I'm not going to do your research for you. Where do you get the information that you are basing your conclusions on? If you don't have specific sources, or if you don't remember them anymore, just say so. As for my source of information, I happen to possess Eliot Higgins book "We Are Bellingcat." In it, he actually opens his introduction with a six page description of...
• Preppy commented on Made of lies (and more lies)
  Well...with AI you need some insight into the technology. Note: Trust in the OWNER of the technology does not count! Technology components: (1) Neural networks (2) Training (3) Large Language Models (LLM) Given the above there are some problems with AI, not all of them technical: (4) Neural networks generally cannot explain their conclusions (but see item #6) (5) Defect #5 gets worse when the neural network is learning AFTER training (6) Large Language Models may (will?) contain an unknown number of falsehoods (7) Only hugely wealthy organisations can own AI tools (Meta, Amazon, Microsoft...). Notably small orgaisations do not...