This is dumb. The solution for all music platforms should be to add a label for AI-generated tracks or artists so users clearly can disambiguate. It's frivolous to prevent someone from enjoying a piece of art whether AI or human. Furthermore, the line is blurred between what constitutes human vs AI development of music. Most producers today use pre-packaged samples, sequencers, and tracks to generate derivatives. Sure, they might manually have to mess around with Ableton to do so, but the line is already blurred.
I suspect the problem is scale. AI slop can be churned out at an unprecedented rate, and Bandcamp is not a big company, but they'd have to host and serve all of that stuff nonetheless.
Solid OSINT methodology here. The 10x AS path prepending is the most interesting detail to me b/c typically you'd see prepending used to de-prioritize a route, which raises the question: was this about making traffic avoid CANTV, or was it a side effect of something else?
A few thoughts:
- The affected prefixes (200.74.224.0/20 block → Dayco Telecom) hosting banks and ISPs feels significant. If you're doing pre-kinetic intelligence gathering, knowing the exact network topology and traffic patterns of critical infrastructure would be valuable. Even a few hours of passive collection through a controlled transit point could map out dependencies you'd want to understand before cutting power.
- What's also notable is the transit path through Sparkle, which the author points out doesn't implement RPKI filtering. That's not an accident if you're planning something (you'd specifically choose providers with weaker validation).
- The article stops short of drawing conclusions, which is the right call. BGP anomalies are common enough that correlation ≠ causation. But the timing and the specific infrastructure affected make this worth deeper analysis.
Would love to see someone with access to more complete BGP table dumps do a before/after comparison of routing stability for Venezuelan prefixes in that window.
I don't think the root cause here is AI. It's the repeated pattern of resistance to massive technological change by system-level incentives. This story has happened again and again throughout recent history.
I expect it to settle out in a few years where:
1. The fiduciary duties of company shareholders will bring them to a point of stopping to chase AI hype and instead derive an understanding of whether it's driving real top-line value for their business or not.
2. Mid to senior career engineers will have no choice but to level up their AI skills to stay relevant in the modern workforce.
the thing is, I tried it and it took 10 seconds to import all settings from cursor. the moat for vscode clones is really small. i imagine people will jump a lot from clone to clone, like from model to model now.
The DRC produces more than 70 per cent of the world supply of cobalt, which is essential for batteries used in electric cars, many laptop computers and mobile phones.
More than 200,000 people are estimated to be working in giant illegal cobalt mines in the giant central African country.
Winter is a cyclical concept, just like all the other seasons. It will be no different here; the pendulum swings back and forth. The unknown factor is the length of the cycle.
I still have to understand why you think another AI winter is coming.
Everyyyybody is using it, everybody is racing to invent the next big thing.
What could go wrong?
[apart from a market crash, more related to financial bubble than technical barriers]
> apart from a market crash, more related to financial bubble than technical barriers
_That is what an AI winter is_.
Like, if you look at the previous ones, it's a cycle of over-hype, over-promising, funding collapse after the ridiculous over-promising does not materialise. But the tech tends to hang around. Voice recognition did not change the world in the 90s, but neither did it entirely vanish once it was realised that there had been over-promising, say.
I suspect he sees a lot of scattered pieces of fundamental research outside of LLM's that he thinks could be integrated for a core within a year, the 10 years is to temper investors (that he can buy leeway for with his record) and fine tune and work out the kinks when actually integrating everything that might not have some obvious issues.
