For the most part yeah but does depend on scale. For Java there are lower cost migration pathways like native compilation anyway if that does become your concern
I think getting personal data collection out of the Thielverse is anything but a strange move
> or if it consumes 2, 3, or 4x the memory I think this used to be true more than it is now. Memory has been relatively expensive in cloud environments for a while (often 2x the price of an ec2 node for an equivalent…
This is actually what keeps people using AI: variable reward schedule. It's basically gambling.
I also find IRL social interactions also want pacing to some degree. But I don't know, I find my brain is sort of wired for these sorts of novel, pseudonymous interactions from a young age.
It wouldn't surprise me. It seems like a mirror of the social media move from social graphs to recommendation engines. You are either producer or consumer and never shall consumer be promoted organically.
I expect many people would accept clicking "license" instead but it's unfair to assume that people don't value ownership as such or that they won't in the future. It's fair to ensure that people are made fully aware of…
I suppose at some point it could have had a wooden structure around it to make it harder to escape
Imo it's a breakdown of trust of the startup ecosystem as a whole. Repeatedly startups have enshittified and it's become undeniable that the investment apparatus around startups is partly responsible. We have seen a…
The other thing to add as well is that the research teams who do the actual research work are relatively small and very specifically qualified which naturally keeps the barrier to entry high.
I think the idea of moral agency in this context is more or less a consequence of accountability or the lack thereof. We don't know a way to punish or remove the threat of such a model when it does something wrong in a…
Yeah this is it for me. What we're really interested in most of the time is whether something can suffer based on observed behaviour. We don't really believe LLMs can suffer so they don't need rights. We believe many…
This seems a lot like the parable of the broken window but for inefficiency rather than destruction per se.
Bit of a tangent but at work we have GitHub Copilot and the VSCode harness is somehow night and day better than whatever happens in the IntelliJ plugin. Aside from having better features, for some reason prompts seem to…
Will it though? Where mastodon is much more distributed, bluesky/ATProto use is highly concentrated on bsky.app
The internet might not be dead, but large parts of it are gangrenous and necrotic
Unfortunately opinions don't change the reality of physical processes.
To be honest the main issue with meta has never been around open/closed software. They've also done react, Cassandra and some other bits. But this, like their open weights is like a feather pressing down on the scale…
I think "code isn't the bottleneck" is often accurate though in organisations with high friction processes.
I think making water expensive because of the fact that it's essential would be problematic at best
I think single dynamic dispatch is one thing but ends up reasonably well optimised by modern compilers, especially the JVM. I think most code written since the 2010s prefers the composition over inheritance pattern for…
I'd be surprised if it were anything other than just an LLM only equivalent of Gemma some 4B param model, they can't be spending much money on it because each query essentially needs to be cheaper than the ad revenue…
The discussion on filler tokens is interesting, but is it not just the case that these filler tokens end up being essentially substituted stand-ins for words we understand with all the same relationships encoded in the…
This one is interesting as it's been hand verified. There was a recent proof that inadvertantly "proved" the collatz conjecture by triggering a bug in LEAN: https://infosec.exchange/@0xabad1dea/117002106099986943
People are already doing it to some extent, especially the sort of technical non-programmer crowd. A member of my family was even having a (now pretty out of date) version of chatGPT spit out VBA macros for them with…
For the most part yeah but does depend on scale. For Java there are lower cost migration pathways like native compilation anyway if that does become your concern
I think getting personal data collection out of the Thielverse is anything but a strange move
> or if it consumes 2, 3, or 4x the memory I think this used to be true more than it is now. Memory has been relatively expensive in cloud environments for a while (often 2x the price of an ec2 node for an equivalent…
This is actually what keeps people using AI: variable reward schedule. It's basically gambling.
I also find IRL social interactions also want pacing to some degree. But I don't know, I find my brain is sort of wired for these sorts of novel, pseudonymous interactions from a young age.
It wouldn't surprise me. It seems like a mirror of the social media move from social graphs to recommendation engines. You are either producer or consumer and never shall consumer be promoted organically.
I expect many people would accept clicking "license" instead but it's unfair to assume that people don't value ownership as such or that they won't in the future. It's fair to ensure that people are made fully aware of…
I suppose at some point it could have had a wooden structure around it to make it harder to escape
Imo it's a breakdown of trust of the startup ecosystem as a whole. Repeatedly startups have enshittified and it's become undeniable that the investment apparatus around startups is partly responsible. We have seen a…
The other thing to add as well is that the research teams who do the actual research work are relatively small and very specifically qualified which naturally keeps the barrier to entry high.
I think the idea of moral agency in this context is more or less a consequence of accountability or the lack thereof. We don't know a way to punish or remove the threat of such a model when it does something wrong in a…
Yeah this is it for me. What we're really interested in most of the time is whether something can suffer based on observed behaviour. We don't really believe LLMs can suffer so they don't need rights. We believe many…
This seems a lot like the parable of the broken window but for inefficiency rather than destruction per se.
Bit of a tangent but at work we have GitHub Copilot and the VSCode harness is somehow night and day better than whatever happens in the IntelliJ plugin. Aside from having better features, for some reason prompts seem to…
Will it though? Where mastodon is much more distributed, bluesky/ATProto use is highly concentrated on bsky.app
The internet might not be dead, but large parts of it are gangrenous and necrotic
Unfortunately opinions don't change the reality of physical processes.
To be honest the main issue with meta has never been around open/closed software. They've also done react, Cassandra and some other bits. But this, like their open weights is like a feather pressing down on the scale…
I think "code isn't the bottleneck" is often accurate though in organisations with high friction processes.
I think making water expensive because of the fact that it's essential would be problematic at best
I think single dynamic dispatch is one thing but ends up reasonably well optimised by modern compilers, especially the JVM. I think most code written since the 2010s prefers the composition over inheritance pattern for…
I'd be surprised if it were anything other than just an LLM only equivalent of Gemma some 4B param model, they can't be spending much money on it because each query essentially needs to be cheaper than the ad revenue…
The discussion on filler tokens is interesting, but is it not just the case that these filler tokens end up being essentially substituted stand-ins for words we understand with all the same relationships encoded in the…
This one is interesting as it's been hand verified. There was a recent proof that inadvertantly "proved" the collatz conjecture by triggering a bug in LEAN: https://infosec.exchange/@0xabad1dea/117002106099986943
People are already doing it to some extent, especially the sort of technical non-programmer crowd. A member of my family was even having a (now pretty out of date) version of chatGPT spit out VBA macros for them with…