Posts

Showing posts from February, 2024

Religion and Trump

Image
WORK IN PROGRESS) I have often explained Trump's success in terms of religious parallels. This is unpopular since it paints an unflattering view of religion itself. There are still parallels, but my research, aided by Perplexity, paints a more complex picture.  We see Trump aligning himself with pre-existing elements of the MAGA playbook. This reminds me of how Hitler used pre-existing antisemitic views along with bitterness over the Versailles peace terms to become a famous voice of discontent. We may gain more insight into the MAGA movement by considering it a classic example of populism . Through this lens, we see the Trump phenomenon as an example of a dangerous trend in history and worldwide. Perhaps we also see something of the future if the Trump movement successfully grabs the reins of power again. Perplexity summarizes Trump's threat to liberal democracy . Here is a defence of Trump , explicitly addressing the belief of his opponents that his followers are delusional.

Fuck Democracy

Image
Trump rallies mob beore the attack on the Capital Jan 6, 2023 Our morning chat today wandered in the deep water when we started asking questions of why people follow Trump. This is indeed the issue. We will always have our psychopaths and narcissists. The danger is that people believe them, follow them, and, at times, are willing to violently attack our institutions in the grip of shared insanity. As always, Perplexity was the third party in our discussion. For one thing, Perplexity doesn't have the safety shackle of GPT-4. It answers questions from the data it has. For example, here is its answer to the question, " Is Trump Crazy? " Of course, he is, but Perplexity sends us down a few side roads to discover this gem from Scientific American , the source of the picture above. This addresses the core question of shared insanity. It also leads us to discover that the question of Trump's sanity is still censored from public discussion by professionals. But I digress ...

Can AI Be Made Safe?

Image
 My answer is, obviously not. Virtually everything from screwdrivers to aircraft transport is "dual use." How will AI be differen t?  Our efforts to keep 12-year-olds "safe" from seeing naked boobies will not prevent anyone from developing autonomous weapons. I will get back to the boobies later. The following answer from Perplexity ignores the use of AI by the military or countries (most of them) beyond the reach of well-meaning chat groups. PERPLEXITY ANSWERS: * Yes, We Can Train AI to Be Safe Technical Research is Growing : The field of AI safety technical research is expanding, with initiatives like the Harvard AI Safety Team aiming to build communities around AI safety research. This indicates a growing interest and investment in understanding and mitigating AI risks 1 . Implementation Frameworks Exist : Google's Secure AI Framework (SAIF) demonstrates that standardized approaches to integrating security and privacy measures into AI development are being de