Why not release the source under AGPL (where 'network use' counts as distribution, unlike GPL), and offer commercial licences for those who want more favourable terms?
The maintainer of libxml2, Nick Wellnhofer, will be moving all his future contributions over to an AGPL fork as corporate users of libxml2 were unwilling to contribute financially.
I pay for Kagi so that I'm not being peddled ads or junk when I'm trying to be productive, as my ADHD-riddled brain can get easily distracted. It gets quite upsetting when I've wasted non-trivial amounts of time on those distractions that I subconsciously fall into.
I absolutely cannot use Google because of their seemingly endless attempts to distract me from what I'm searching for.
The final nail in the coffin was their actions to get rid of uBlock and other effective ad-blockers. It's a serious anti-pattern, and (I strongly argue) is effectively discrimination for those who struggle with ADHD.
I hope that Kagi can one day effectively filter out GenAI slop websites that look like legitimate content, but I can understand the significant technical challenges in such a feature.
I can't say I'm surprised by this. The brain is, figuratively speaking, a muscle. Learning through successes and (especially) failures is hard work, though not without benefit, in that the trials and exercises your brain works through exercises the 'muscle', making it stronger.
Using LLMs to do replace the effort we would've otherwise endured to complete a task short-circuits that exercising function, and I would suggest is potentially addictive because it's a near-instant reward for little work.
It would be interesting to see a longitudinal study on the affect of LLMs, collective attention spans, and academic scores where testing is conducted on pen and paper.
It's like a drug. You start using it, and think you have super powers, and then you've forgotten how to think, and you need AI just to maybe be as smart as you were before.
Every company will need enterprise AI solutions just to maybe get the same amount of productivity as they got before without it.
And the pipeline is cooked now with some universities now allowing for AI use. It’s like what cliffnotes did for reading comprehension but over all aspects of life and all domains. What a coming tsunami.
Most speeding offences require the use of a speed measuring device to detect and 'prove' an offence. However, a number of jurisdictions have a separate offence where 'speeding' can still be charged, including 'in motion', without lidar or radar.
For example an officer following or pursuing an offender can apply a 'negligent' or 'wreckless' driving charge based in context of the officer's observations and evidence gathered, such as following or pursuing an offender well above the speed limit, observing the calibrated speedometer in the patrol car, without the use of a speed measuring device.
It's been a while since I've looked at it though some Australian police forces have a calibrated speedometer installed on the dash that reads out the vehicle's speed based from the rear differential[1], separately to the vehicle's 'stock' speedometer. The reasoning, I understand, is that this is more precise, as legally the stock speedometer can display a speed up to 10 km/h lower than actual (but not above).
It's the other way around: the speedo can overestimate your speed but not underestimate it. If you follow the limits with an overestimating speedo, you drive under the limit. With an underestimating speedo, you end up over.
Anecdotally, when I pass those roadside speed alert signs, the speed they show and the speed on my speedometer is rarely more than +/- 1 mph. I think modern speedometers are pretty accurate, as long as the OE tire size is used.
That my experience too, the speedometer, speed my phone thinks I'm going, and static radar signs all more or less agree. Plus the only times I've been nailed for speeding, I was speeding, not "just kissing" the limit. Point being I don't think +/-1mph really matters in practice 99%+ of the time, it's usually getting tagged when being overconfident in the passing lane or something like that.
I would definitely use this to charge US$100,000 per request from any AI company to crawl my site. I would exempt 'public good' crawlers like The Internet Archive though.
If AI companies valued at billions of dollars want to slurp up my contribution to the human condition, that's my price - subject to price rises only.
It's really shitty that companies believe they can pull these stunts and get away with it.
reply