One of the 2 names on here says they aren't being listed as an author on the paper because their contributions don't meet the standards of authorship. People can call LLMs 'tools' or whatever they want but if you had essentially nothing to do with the breakthrough, you're not an author.
What country are you in? Not challenging, I'm gathering perspective.
I am in the US with health insurance through my employer and I get blood tests every 6-12 months to check for abnormalities as well as a standard interview for how I'm feeling and whatnot. Also have a few prescriptions that require 3-month checkins.
As I understand it, Mojo’s goal is to be able to match state of the art performance for parallel algorithms and be able to use the exotic GPU hardware. A functional style is not necessarily productive in that context.
1. Most of the negativity about AI seems to relate to the economic effects, not the technology itself.
2. From what I've seen, the average person is still unable to identify AI created content, websites, etc. The website this post is about is clearly vibe coded (looks like everything including the copy is from Claude) but I'd guess that only a small fraction of the people who see it will recognize that.
Maybe not organised like the mafia but you 100% get crackheads shoplifting in an organised fashion to get money for drugs in London, and self checkouts obviously make this easier.
Not unheard of to get crackheads trying to sell you stolen supermarket meat out of a bag if you hang around in some of the less salubrious pubs
Don't want new features, and don't need an upgrade then stick with it. If you need the new version, then pay a partial amount of the app for the upgrade?
Looks like your great question doesn’t have an answer, but looking at the routing strategies things get even more confused, since the proposed ones tend to rely on extra llm calls to determine which model to pick.
The nice thing is that it makes sense for specific setups, less conversation oriented.
As an example, you need to classify batches of data, and have many fine tuned models. Or you need to do speed to text and need to pick which whisper to use.
You can write your own strategy, in that case an harness with subagents would be able to leverage this, picking the right model and then keeping its session sticky, but overall the lack of concern for caching points towards use cases where you do not gain much from it.
You expect every blog post to find the earliest relevant paper to cite, just so one could look at the year (without reading said paper - which would have made clear that the connection isn’t recent) to assess novelty? I don’t think that’s reasonable.
It’s a blog post. If it was, say, a peer reviewed paper by Hinton or LeCunn that fails to cite Schmidhuber, that would be reasonable criticism in my opinion. (Spoiler: they fail to cite him)
Not my experience, I've found it extremely useful for when I need a quick answer. It's handled very specific fairly obscure questions well. I also find it quite consistent.
If you haven't yet, you might want to try gemma-4-E4B-it-qat or gemma-4-12B-it-qat with Structured Outputs. My main use case is tagging photos and generating headlines.
I've found their models to be weaker and their inference much more unreliable, which is unexpected since they have their own TPUs and you'd think this is something they can handle. But no.
The contradiction is the notion of having hardware to support metaphysical dualism. It's an oxymoron.
I think what they're actually trying to say is that the physical reality of LLMs doesn't leave any metaphysical space for dualism, which is required in their view for "mind". Ed: or maybe not, based on their other comments.
