Even if that happened tomorrow, I suspect we'd have _at least_ a decade of people tweaking/optimizing designs on the same node to squeeze meaningful performance upgrades out. Eg, coming up with hardware support for new int/float formats that make more sense for the models of 2029, running matrix operators on ram chips directly, etc.
I remember back in the early 2000s when people thought we were running out of steam on the advancements front. This was roughly around the time when CPU clocks stopped getting faster. Pentium hit 3GHz in 2005, Intel Core Ultra 5 performance cores are generally around this exact speed 20 years later.
Since at least the 640kb quip, betting against progress or the appetite for progress has been a losing bet.
Honestly post 2005 things did slow down dramatically for typical single core workloads.
In the late 90s and early 2000s the mantra was "why waste time optimizing your software? By the time you're done the next gen of CPUs will have made up the difference."
Now the increase is more about moving to GPUs and power efficiency etc. We still have increases, but the rate of speedup has slowed down a lot.
For the core of his base certainly, but he can't win with just them, it's not enough. He won because of the median voter that was mad at the incumbent party for inflation.
> They haven’t been able to fix it with many multiples of that.
which may actually be the problem. I suspect that there is actually some ideal ratio that could be calculated of Input Fields / Dev, LoC / Devs, or maybe Unique Pages / Dev, or some mix of all of the above. Some of the metrics I hear out of places like airbnb absolutely blow my mind (>5000 engineers! wtaf are they all doing?!?). I can sort of see the #s at google, MS making sense given the breadth of the problems they are solving, but other places, not so much.
There was an interview with a lead tech at Uber about this some years ago, with the conversation starting with "why is the app so bloated!?" (in terms of megabytes) and his answer also answers your question:
The smooth and simple interface of the Uber app is the tip of the iceberg. His example was that their users don't (and won't) reinstall the app because they travel overseas. If anything, the time when you've just stepped out of an airport is precisely when you want the app to work smoothly!
The hiccup is that many countries have their own payment systems, Byzantine tax codes where this may or may not be displayed up front to the user (in various currencies and formats), there may be local laws around taxi-like services, etc... Some of those laws apply to areas smaller than a city, or may apply only to airport pickups, or the CBD area during congestion, so on and so forth.
The "core" app might be a simple thing that you can bang out over a weekend with an AI and a decent UI framework, but then you need to "draw the rest of the owl". Don't forget that there must be a matching app for the drivers! Different categories of drivers offering services that may be local to a region and totally absent elsewhere: rikshaws, tuk-tuks, taxi boats in Venice, and who knows what else!
AirBnB is very similar to Uber in this respect. They have to deal with about a hundred countries worth of law, often down to the state level. There's fraud detection. Customer support. Integration with travel agencies. Government-mandated reporting. Etc, etc...
Project size is obviously going to be a factor, but so is machine specs. It's much more noticeable on a spinning disk. One can partially compensate for the project size aspect by opening vscode as far into your project as possible (eg, the api subfolder) rather than at the root. No real solution if you don't have an ssd though.
Because of how policies are grouped together in many countries (particularly the US), the fight against fascism is necessarily concurrent with the fight against climate change.
This is a function of your personal media diet: when you spend time consuming rationalizations funded by the fossil fuel industry, it’s easy to deride people who were right as “screamers” and the longer you do so the easier it is to project your intellectual failure onto nebulous “other” people you don’t know. It’s not your fault that climate change is costing lives and billions of dollars, the people who tried to stop it were engaged in hyperbole. It’s not your fault that people are being illegally killed, assaulted, or detained, it’s their fault for allowing you to ignore the clear evidence that this was going to happen.
I've been reading the news accross views and geographies for a looong time. Many people would say things are going great with a lot of long-standing problems getting solved.
"Many people" "a lot of long-standing problems getting solved"
Weasel words. How many? Who, specifically? Which problems have actually been solved and aren't just "solved" via people metaphorically standing on an aircraft carrier with a "Mission Accomplished" sign pretending they've been solved?
In my social circle of liberal people, the reason is despondence.
Climate change has been known for decades now, and despite the alarms and concerns, the current administration is cheerfully, maliciously removing all initiatives in the US to combat it. Attempting to destroy the solar industry and wind power. Rolling back the most common sense environmental controls for public health.
Meanwhile our country has had its place in the world destroyed irrevocably (for at least a generation) and is turning further and further away from a country that cares for its citizens and its freedoms.
People are losing hope, not interest, because climate change and fascism are are more alarming than ever and our government is complicit.
But not 38 TOPS that Apple claims, with the weak explanation of
> Apple’s “38 TOPS INT8” is computed as 19 TFLOPS FP16 × 2, following the industry convention of counting INT8 operations as 2× the FP16 rate. But the hardware doesn’t actually execute INT8 operations twice as fast.
Why would Apple follow that convention when the hardware explicitly doesn't seems like a more straight-faced lie that I expect from Apple
there's an apocryphal story that when one of Apple's chips was nearing 10B transistors, marketing asked the chip folks if they could round it up to 10B for their copy. The chip folks were confounded, and said no they didn't have any uncounted transistors to round it up, and they didn't approve of claiming 10B transistors when it wasn't.
(This was a while ago. I see the M4 is at 28 B)
Which is why I'm all the more surprised that Apple would claim 2x more ANE TOPS than it can really does.
That's not clear to me, and apparently not to Block.
> There’s definitely going to be a shift over time.
Yep, and as the transition happens, it'll probably be bell-shaped. Someone will be on the left side of the bell, and someone will be on the right side. Block has chosen to be on the left side.
This feels like a very, very weak prediction (though certainly possible).
reply