No phones can run “LLMs” currently because by definition, large.
Some Android phones however can and does run smaller models locally. Gemini Nano runs on Pixel 8 and can run on Samsung phones.
No phones can run “LLMs” currently because by definition, large.
Some Android phones however can and does run smaller models locally. Gemini Nano runs on Pixel 8 and can run on Samsung phones.
It’s not a LLM, it’s a much smaller model (~3B) which is closer to what Microsoft labels as a SLM (Small Language Models, e.g. MS Phi-3 Mini).
https://machinelearning.apple.com/research/introducing-apple-foundation-models
There is a lot to unpack from your post. First of all, there is no doubt that technology in general adds value for the human race - like the another commenter said, foundational things like fire, tools all the way to the zipper and buttons you have on clothings, umbrellas you bring into the rain, the video chats you have with loved ones during COVID - those are all the fruits of technology.
But if you get down to the particulars, value can be very subjective. Some people value fancy new tech sneakers, primate NFTs whereas others value new computer vision technology or a new programming language. So are certain technologies adding value? Depends on who you ask.
As for who is capturing value in a capitalistic society, I think you already have the answer. Simply put, if your company operates at a 50% efficiency and you bump it up to 70% with tech and automation, be assurance you are going to see job cuts to " and increased targets to produce more - you are not going to get more leisure time but instead be asked to push ahead until you hit the physical limit and break.
Fai point, but regardless it seems to have struck a nerve with the piracy crowd.
I don’t have beef with piracy itself but I found it hilarious the number of pirates here standing on their soapboxes, pretending to be some kind of modern day Robin Hood and virtue signaling super hard.
Guys, you are still ripping off artists and content creators regardless of their deals with media company, just admit you want shit for free.
The amazing mental gymnastics that these people go through to justify their piracy and inane behaviors.
Musician’s pay is just the excuse of the day for them to feel okay about what they’re doing. Honestly, if you are gonna pirate then just pirate, stop pretending that it’s for a good cause or higher purpose, other than to keep your own wallets stacked.
It is not anti-woke, which by anti-woke logic means it’s woke.
it’ll go along and do it for a second, but then it’ll quickly wipe its message. That’s interesting because it suggests that the underlying model isn’t the same as OpenAI’s
Obviously I can’t speak to Bing’s claim on GPT4, but the behavior you saw does not necessarily have to do with the model. There are many ways the “chatbots” on the same model could behave differently, either by defining those behaviors programmatically or via context and prompts.
I’m not sure why so many are rooting for Beeper. Apple’s response is 100% reasonable - you have a 3rd party service that’s making money by impersonating iOS devices in order to access Apple services. Apple has no way of controlling how many devices will use Beeper and if their system can maintain a good level of service, how these Beeper devices are interacting with iMessage, and whether Beeper is actually keeping iMessage metadata private or just giving lip service.
An analogy would be like Apple is throwing this awesome concert event and Beeper found out a convincing way to fake the tickets, and are actually actively promoting, registering people and profiting off of it. In any reasonable world outfits like this would be shut down immediately and rightfully so.
Short version: yes, Sutskever was a part of the coup to kick Altman and Brockman out. What happened next was Brockman’s wife pleaded to him, then he changed his tune afterwards.
Not for me
most of what you said sound pretty good ngl
The product Beeper is selling is essentially access to “blue bubbles” and iMessage without having to pay the price of admission (i.e., owning an iDevice and working with Apple). That’s the part that’s shady and sus. What they are paying or saving on infrastructure cost is irrelevant - they are basically still running a counterfeit operation, doesn’t matter what their costs are.
The tech is pretty interesting but the business is sus…It’s kind of like selling fake admissions to a club and calling that a startup.
…Beeper CEO Eric Migicovsky responded to TechCrunch’s inquiry about Beeper Mini’s status by pointing us to the X post acknowledging the outage, and providing more detail. Asked if possibly Apple found a way to cut off Beeper Mini’s ability to function, he replied, “Yes, all data indicates that.”
Emphasis mine. Source: Techcrunch.
make it available to anyone
To do what?
I mean, if there are people who want to buy it, why not? It would just be icing on the cake for Bethesda.
but it’s not like they add it for no reason.
I didn’t say anything about that. I’m saying the main reason Bethesda removed Denuvo from Doom Eternal is likely because of cost reasons, not because it’s a marketing play to drive sales (like OP suggested).
Sorry but that doesn’t really make sense. In that scenario it is more sensible to just release a DRM free game at start, because the first group would buy either way and the second group would buy at the higher launch/near-launch pricing (since games drop in prices over time). It doesn’t make sense to make essentially 2 versions of the game over such a span of time like you described.
A more realistic scenario would be that there is some cost / licensing fee to use Denuvo tech and it no longer makes financial sense for Doom Eternal to do so, hence BOOM! DRM free.
I have been running a TP-Link Archer C4000 for several years now. Super solid, great price, no complaints. For me the tri-band capability is more important than Wifi-AX, which isn’t very useful unless you need incredibly fast network speed or if you have a ton of devices (which all need to be supporting AX).
I do think it’s a useful distinction considering open models can be more than 100B+ nowdays and GPT4 is rumored to be 1.7T params. Plus this class of models are far more likely to be on-device.