"Intrusive" may be a C++ speak, but I wouldn't say that one or the other type of lists is necessarily much older or more "normal". After all, a cons cell embeds data and not the other way around and lisp is one of the oldest programming languages in existence
What makes the front wheel brake a bad idea? I almost exclusively use the front brake, because that's the one that'll actually stop me. (Source: when I used rim brakes I had to replace the front brake pad three times for every time I replaced the rear brake pad.)
I know some people use the rear wheel locking up as a way to sense when they have applied enough front brake, but that's not necessary. It's also possible to learn to sense that directly.
When going at a high speed, breaking hard with the front wheel alone can force the bike to rotate around it and the rider might enjoy a very short but very thrilling flight. Which I learned the hard way many years ago
It isn't, the cost is included in your electricity bill, not even talking about the cost of your time to set it up. It's very possible that it costs you more than a cloud mode would, you just don't want to calculate it properly.
It sounds like they are doing something similar to what I described in my other post below. Personal media station.
That can be done on hardware that quite a lot of people basically just have and don't use 24/7 to the max - because it is their gaming machine or their programming and compiling workhorse, for example. Of course you are paying for additional electricity but even with napkin-math instead of a "proper" calculation, you are unlikely to pay more for running your own instead of something commercial (and that can be offset further with some of the "modern" electricity contracts and/or PV and battery storage). Especially if we are talking about a stack that runs most of/all the time when you are not using your machine and makes LLM calls regularly while running.
The work in software/admin to get whatever you want set up is similiar no matter which infrastructure you use.
true but if you're actually running k8s and similar workloads, chances are it might eat memory that LLM requires.
you'll also notice these articles rarely specify their context window in tokens, because it is small, usually 30k to 70k tokens and it gets slower as it fills up.
I actually have a Mac Mini M4 Pro with 48G. I gave the k8s example because this is what I was doing with it.
Was because I am back to using Linux as my workstation.
My Mac Mini is now a headless server for llama.cpp.
So, you are right that for these workloads , I would not be using the Mac Mini for k8s AND llama.
Another thing going against using a Mac for Linux containers is that there are no solutions that I know that properly manages memory : memory is given to the Linux vm , but never fluctuates if the needs in the vm are less than the initial request.
I know Orb Stack does that but is it proprietary. I think UTM does it , but not sure I would use UTM instead of Lima, Colima , multipass , etc to run containers.
> It's very possible that it costs you more than a cloud mode would
...which is almost always true in a single request/reply mode and never true in batch mode. Single request usually 2x-3x more expensive than cloud and batch mode 2x-3x cheaper. Now, for narrow tasks, a finetuned tiny 8b model would dramatically outperform SOTA frontiers for a fraction of price, esp. on energy efficient hardware like Apple.
Local is never cheaper than cloud because they can do batch inference, and that means you load model weights once to produce 128 tokens on 128 sessions in parallel not 1 token on 1 session like local models. Local models rarely get to high utilization factor, they spend most of their time waiting.
If you had only batch inference and enough of it to fill the compute to 80% then you get cheaper local models.
Local models can absolutely run in batch, what are even talking about?
> If you had only batch inference and enough of it to fill the compute to 80% then you get cheaper local models.
Even if you ran sequentally, single session, a _finetuned_ tiny (8B) local model on narrow tasks would abolutely mog SOTAs, any of it - Fable, Opus, Sol you name it.
> Local models can absolutely run in batch, what are even talking about?
I think the point was that if you aren't running your local machine at 100% for 24 hours a day then a cloud - with multiple clients - that is, will be more efficient.
I am so envious of DC and Chicago for their metro systems. It's always interesting to talk to people who live in those locations and take their mass transit for granted, like a spoiled entitlement. Here if you don't have a car you will quickly become homeless or starve to death.
If Google is right they live in Jacksonville, which aside from being a terrible place to live in general, can't physically or geographically support subways or light rail. So you're basically limited to busses if you want to implement public transportation.
I've only been to Jacksonville a handful of times but I'd rather walk along the side of road in August than ride a bus there. And they don't actually help congestion that much.
Jacksonville (and most of Florida) is extremely low lying with a high water table, so sub-surface construction (basements) and subways are basically impossible.
Structurally, it's a low density, car-based city so light rail (as opposed to busses) requires a lot of up front investment without the land use to justify it. At this point, like many US cities, adding effective transit is a chicken and egg problem with land use.
There are a couple reasons (not the least of which is political will but that's a separate discussion). Jacksonville sits on a coastal plain and is almost entirely fine sand. You simply can't build a subway in that without astronomical cost. As for rail I'm pretty sure Jacksonville already has elevated rail in some places but ground rail would be difficult for the same reasons.
I would like it to pick up shoes, clothes etc. Not sure the beak would work, it would block the camera and prevent the bot from walking forward. Presumably, it could walk backwards and drag clothes, but anything rigid might block the camera.
yeah they still have the ticket stations next to the one to refill/buy IC cards. and thanks to the 20/21 Olympics in english now too(not 2 levels of menus deep like many of them used to be) so pretty easy to do
How much of it is "the current administration" and how much is simply the continuation of a much longer trend? The US has been at war with this or that country in the Middle East for as long as I'm alive and I'm almost 40. Same with tax cuts, high spending and so on. It's easy to blame everything on Trump (or Reagan), but it seems to be a much broader issue with the whole system
Yes, this has been going on a long time. However, I think there is quite a bit of data about who is in charge or the White House and Congress when taxes go down/up or the deficit goes down/up, etc. It doesn't take more than one online search to see a very distinct pattern.
Democrats Change in Deficit
--------- -----------------
Joe Biden -6.1% GDP
Barack Obama -6.4% GDP
Bill Clinton -5.0% GDP
Jimmy Carter -0.2% GDP
Republicans Change in Deficit
----------- -----------------
Donald Trump +8.6% GDP
George W. Bush +11.0% GDP
George H. W. Bush +1.1% GDP
Ronald Reagan +0.2% GDP
Gerald Ford +2.3% GDP
Richard Nixon +0.7% GDP
Dwight Eisenhower -1.1% GDP
Let's count our blessings that the ongoing campaign to fill whitehouse.gov with Republican partisan hackery doesn't extend to the boring accounting parts. Yet.
And yet every single Fox talking head says the opposite, resulting in half the country thinking it's Democrats responsible for the deficit. Propaganda is magic.
Thank you for this. I am so tired of false equivalence. Democrats passed Obamacare, the chips act, the infrastructure bill and countless other large impactful pieces of legislation that people take for granted. And did that while reducing the deficits. Republicans killed investment in solar energy ceding the industry to China, same with EVs. We would be better positioned with abundant electricity with AI needing vast amounts, but we’ve put them in charge yet again at a critical moment. Biden had a coherent policy that would have been predictable. Trump randomly banned fable but not gpt5.6 which caused most the world to decide that reliance on American AI is extremely risky…good job.
Hehe its not that complicated, if it were, Trump wouldn't have bothered taking on the job haha according to the school of thought I've gathered from conservatives, its "loot or be looted". Not much cognitive effort needed here, "extremely complicated" lol u can miss me with that dodge
Clinton balanced the budget. There was a surplus. Go read
Articles from the 90s there was fear that we would pay down the debt too fast and it would harm the economy.
Obama was on tracy to reduce the deficit and likely would have had the financial crisis not required stimulus.
Biden was using debt to invest in infrastructure and things like power generation that would yield returns unlike tax cuts that just exacerbate wealth concentration and inflation by asset appreciation due to capital chasing the same assets.
It’s not symmetrical one party is driving this and the irony is that they are the “fiscally responsible” party.
No amount of discipline is possible for democrats if every time they balance the budget republicans use the savings to cut taxes further.
Printing hundreds of billions of dollars is also inflationary. Governments change but the impact of all this takes a very long time to flow through the economy.
Except American was in a structural slump similar to but not quite as bad as japans lost decade which meant that interest rates were stuck as near 0% percent. The inflation caused by Covid actually broke that cycle and American was on a much better path. Go look at articles from the last year of bidens term in the economist and financial times. The US had achieved a soft landing by bringing inflation down to within fed targets while avoiding a recession. We were the envy of the world. Income inequality was reducing, real wages were increasing for the first time in generations and productivity vs wages were converging.
Now we’re back to chaos and inflation and wage stagnation.
Yep. Biden handled this gracefully. And i can confirm internarional envy.
The part gggGP doesnt seem to grasp is, the fiscal hawks seeking to cut taxes and regulations etc. are strong and old lobbies around both parties and thats why you can see a trend. Some elected officials are just less susceptible than others and trump is a giant jackpot for the parasites, but so was reagan back in the days.
foreign policy has always been a mess, and the US has always been a bully
but in terms of the deficit, the difference between Trump and other admins (Bush aside) is stark: just take a look at the table in one of the other comments (which I won't repeat here)
Most definitely. Trump just added a little bit of gunpowder to a fire that was already burning pretty self sustainably. Just a little turbo boost to get to the dark future a bit quicker.
If the fire wasn't already burning, Trump wouldn't have been able to bring such a number of cronies together - just look at the tech bros, they already had everything, but they still chose to lick the boot because they want moar!
