author is describing the experience of someone with poor prompting skills. when my prompting was poor, I felt the same way. It’s GI-GO. I worked hard to improve my prompting, using a wider range of subject matter, and got a lot better at using the tool. I don’t have that experience ever now, and it was all about becoming a better tool-user.
Is it ‘harming democracy’, or merely exposing a system of theatrics & ceremony intended to masquerade as a ‘democracy’? I’m less convinced our democracy was ‘lost’ than I am its absence was long-obscured by clever elites.
The counter argument is an expansion of ‘no man is an island’. Everything you might try to liberate as ‘yours and yours alone’ is simply a fragment of ‘The Commons’ that many before you have created and enabled, which you are now attempting to temporarily appropriate for your personal discretion. Sovereignty of an individual is a contradiction of ‘society’, not a facet of it, and its pursuit merely an unambitious form of plundering.
I am uncomfortable with this view yet it is challenging to rigorously defend against. Well-considered attacks against it are welcomed.
More guidance is needed on how to efficiently prompt an LLM to generate a correct UI. Describing the appearance, placement and behavior of UI elements for this is quite time-consuming. The contrast with server-side code generation is stark.
I agree with your observations, but my conclusion from it is therefore karma isn’t worth worrying about. Not uncommon for me to see heavily upvoted material I disagree with, or vice-versa, and am resigned to the fact that bot traffic will be a significant factor on both counts. I still get to view both. Unlike reddit, it seems unpopular opinions still have a home at HN. Karma is mostly for the bots, not me.
it’s joyless, but AI seems great at old legacy software integration and sustainment. Companies have old packages they cannot move off very easily/quickly, lots of legacy data, and AI is very good at coming up with creative ways to sustain/integrate them, and in the process automate the most skull-numbing/laborious/high-risk aspects of that work. (I’m dealing with +20yo packages right now, doing things I NEVER would have agreed to even attempt in a pre-AI world.) Proprietary data formats, obscure protocols, nothing is a show-stopper anymore. No fun, but pays the bills.
A ‘scheduled transmit’ could be an effective & simpler alternative. Some email clients allow you to compose a message and then schedule transmittal at a specified future date/time. Something similar for txt msg could be useful.
The ‘cover story’ is that he ‘warned Russia not to attack NATO’. What they actually talked about was certainly different from, but possibly related to, that talking point. Cover stories serve useful purposes, that’s why they are promoted.
I eat oatmeal for breakfast 5 days/wk; add a handful of chopped walnuts and some apple pie filling. The cans of pie filling are inexpensive/simple and the combo makes oatmeal easy to eat regularly.
Combo of reputation & provenance. Someone widely respected as a super-competent programmer and known to be somewhat of a self-directed loner. If Fabrice Bellard suddenly decided secure messaging was what he was committed to doing, I wouldn’t care if he did it closed-source.
Another vulnerability is natural gas pipeline compressor stations are all internet-connected, often via wireless because they are in remote locations. Simple DOS attack against one will force it to shutdown flow, causing things like power stations to drop offline.