Specialized research AI agents are coming at which point we'll have numerous LLMs running and verifying experiments and creating a higher quality text corpus than the 2014-2020 halcyon, which is then used for other LLMs to be trained on.
It will be the reverse I suspect. Eventually we will see that LLM quality is lower when it is training data from 2014-2020 and will chalk it up to human limitations and the data not being written with a laser-focused goal of training better AI.
I feel like all the BS we were taught about architecture design principles multi-AZ, failover strategies, graceful degradation etc was gaslighting us all into thinking any of out work on it actually matters.
This isn't true, but it feels like this when the entire engineering world order seems to actually run on single-point-of-failures where one CEO just messages another when some 3rd party is down. And reputational risk here is completely safeguarded because as long as everyone is down you are fine. Use a service everyone uses and it goes down = no reputational risk. Use a more robust architecture and make some mistake = massive reputational risk and everyone asks why you don't use what everyone else uses.
I just want to point put there that your argument's exact same rhetorical structure could be (and has been) used to deny "rape culture":
E.g.
- "Rape is illegal and prosecuted, so how can we have a 'rape culture'?"
- "That's not rape culture, that's just individual bad actors"
- "People criticizing women's clothing choices is normal social interaction"
- "Rape culture is a partisan feminist concept like [insert dismissive comparison]"
The parallel is that both involve:
1. Demanding an impossibly narrow definition (complete silence vs. systematic legal tolerance)
2. Dismissing patterns as "just normal social behavior"
3. Focusing on whether the most extreme version exists rather than whether there's a meaningful phenomenon worth discussing
4. Using the term's political associations to avoid engaging with the substance
The irony is particularly sharp when you argue that "telling someone to shut up" is quintessentially social while simultaneously arguing that coordinated efforts to damage someone's reputation/livelihood for speech don't constitute a distinct social phenomenon worth naming.
This was a once unscripted statement that she never repeated again her entire campaign. She expressed regret immediately after making it (not after losing) explaining how it was grossly generalistic but stood by that it does describe a large chunk of Trump's base.
Can you name one time Trump immediately regretted what he said about the Left and explained that it was an overgeneralization? Also again, in Clinton's case it was a single unscripted statement. By contrast Trump has a pattern of calling opponents vermin, the anti-christ, evil, warning of a bloodbath to come... Yea the left's attacks on the right just don't seem nearly as bad especially if it is just being called "deplorable" once. I could write a book that is just hateful Trump quotes. Ganna print it and give it to my son one day as inspiration for how to be an honourable leader. /s
I say all this as a Canadian that would have voted for Trump in 2016 if I could have.
> You check your banking apps multiple times each day with the frequency and unpredictability expected from messaging apps?
I almost do, yes. Life's complicated, I use several bank and credit card services a day. I'm not at home at suitable times for my banking needs. And payments for purchases sometimes require confirmation in real time via phone app.
> If not that frequently or unpredictably then you could just plan to use your laptop for banking some time during the day.
I used to do that years ago back when it was an option.. But these days, 3 banks I use (two for business) require using their mobile app to authenticate login on a laptop browser. There's no other option.
One of the card apps I have to use often won't even run when Android developer mode is enabled, which is quite annoying.
