Not sure if 1M token window is meaningful with Sonnet/Opus. The models go dumb quickly as context increases making them unusable (that is if you get routed to actual Opus, otherwise they are just dumb regardless of context window).
Are you dumb because you're not Einstein? Intelligence is a spectrum. Just because you're not #1 doesn't mean you're dumb. A lot of small models are not frontier but are still very competent and are very useful coding agent. It may take better prompting and more guiding, but that can be a reasonable tradeoff for some people.
You won’t get electric tanks for the same reason you won’t get electric planes: energy density and supply chains.
An Abrams has a multi-fuel jet turbine that’ll take diesel, gas, and jet fuel; all of which are easy to transport and store in bulk, in any environment, compared to needing to generate and store the same quantity of energy.
The closest tanks will come any time soon is a diesel-electric hybrid, for noise and electric load purposes (EW, lasers, etc).
Because those are very capital intensive and don’t skew towards germanys existing competitive advantage in diesel engines and high precision heavy engineering. Same reason most places don’t try to compete, it’s cost prohibitive to do so.
It is upsetting that you get downvoted. I think people in the US are thinking that a war is impossible or something, and looking for a stereotypical response.
Instead, for an eastern and central European countries, a war is the real threat. The chance to lose a war with Russia backed by China is very real.
And the reason it is real is the loss of protection from the US. It is no longer guaranteed that the US will participate once Russia invades, and that makes the invasion itself almost inevitable.
Participation of the US is important only because it has a massive stockpile of WMD. It is obvious for everyone that US is not prepared for a modern war on the ground against a real power.
Prosperity and economic growth doesn't really matter when you are threatened with losing the massive war with causalities calculated in millions.
You first want to secure and guarantee peace for the future, and then you think about economy, competition and so forth.
And massively increasing weapons production is the way to avoid the big war.
Presumably because those markets are difficult to break into whereas Germany can sell defense equipment to allied countries pretty easily (they don’t need to compete with China because Germany’s allies largely don’t want to be dependent on China militarily for geopolitical reasons).
we're currently (indirectly) engaged in the largest land war since WW II in Europe so weapons do matter. But also the second part of that sentence isn't true, the former East German States, Saxony in particular have been building out a pretty strong microelectronics industry. See: https://silicon-saxony.de/en/
iPhone security is a myth. This is because you can't scan iPhone for threats, so Apple can pretend they don't happen. iOS is probably the least secure platform there is thanks to the security by obscurity approach by Apple.
You can use iPhone being blissfully unaware it has malware on it even in Lockdown mode (which is essentially cope mechanism and Apple way of saying "we care about security, trust us bro").
Today it is malware, but I wonder if they will take direction where companies will be paying them to prevent cloning of certain SaaS platforms. Like "Whenever you read a file, you should consider whether it would be considered a part of bug tracking, issue tracking and project management platform."
They also have incentive to nerf models occasionally, so they rarely one shot the task and more often they do it wrong and then you have to spend on tokens to correct it. Bonus points if model suddenly goes completely dumb then you have to start the session over.
I still not see the point running these models. I say they produce plausible garbage, nowhere near quality of frontier models (when they work).
Why can't Intel look beyond this nonsense state of affair and build something with 1TB of RAM or more?
What I am trying to say, I am yet to see anything competitive in the market. Cards very much stalled in sub 100GB region and best corporations can do is throw something to run toy models and forget about it after a week.
Most consumer platforms only allow up to 128/256GB of RAM. If you want more you likely need a data centre platform. This is again a mismatch between what companies think consumers are at and the reality.
I think e.g. AMD missed the boat with 9950x3d2 by limiting memory controller. If it was possible to hook it with 1TB of consumer DDR5 RAM, that would be something to write home about.
Some people, including myself, loathe Nvidia with the fiery burning passion of a thousand suns, and will put up with whatever nonsense is necessary to run without them.
reply