I don't think sticking it out was optimal - there's no airflow past the heatsink. Another mate overclocked and left the case side off, a big desk fan was contantly directed at the motherboard - seemed to do the job of keeping temperatures of his dual CPUs (differently cloked lol) down, was obscenely noisy, but for determined maddened gamers, I suppose it's a small price to pay.But would that expose the PCB?
there is no ai it is machine programming from data collected if it had free will we would all be screwed.But would that expose the PCB?
People are still trying to push the hardware instead of AI or architecture. My sony tv does picture enhancement so well that I rarely if ever select channel 101 instead of 1.
My recent experience of managing to repair my freezer using superglue + baking soda, causes me to believe any shape or size PC case is possible to cut and glue together using old cases. It's bloody strong stuff! Then there's the even harder superglue + graphite to try.problem is if i went diy id have to make it work in the case i want with a custom heat sink and cooling
That technology can be faster of course, when baked into dedicated chips. This runs on general purpose:People are still trying to push the hardware instead of AI or architecture. My sony tv does picture enhancement so well that I rarely if ever select channel 101 instead of 1.
I hate that the DJI Osmo Action 3 has to be activated by an app - IT WON'T DO A DAMN THING WITHOUT ACTIVATION! This has almost completely put me off buying one, but it looks more reliable and better overall than Insta360 and GoPro offerings. The Chinese gov will have my number and network access...there is no ai it is machine programming from data collected if it had free will we would all be screwed.
0 carbon=
How could AI help speed up general purpose computing tasks if built into CPUs? Would have to be aided by ML in the cloud surely? Or have local access to large pre-learned models and datasets, updated/synced by cloud periodically?People are still trying to push the hardware instead of AI or architecture.
I was sceptical about AI for a long, long time. The early 'neural nets' in the 90s were pretty rubbish. The case for AI is now well established. It's like 20th century physics, until Einstein. You could have worked out that E=mc2 before Einstein introduced Special Relativity but not what electronics that was born from this new branch of physics, can do for us. We are still in the very early days of AI, eventually, we'll have to trust the machines because we can never have many good attributes that machines have.How could AI help speed up general purpose computing tasks if built into CPUs? Would have to be aided by ML in the cloud surely? Or have local access to large pre-learned models and datasets, updated/synced by cloud periodically?
And therein lies the problem.We are still in the very early days of AI, eventually, we'll have to trust the machines because we can never have many good attributes that machines have.
we are still many, many years away until AI becomes 'sentient' and can be trusted with real power.And therein lies the problem.
True AI, if ever reached, won't take long to realise that the only real problem to be eliminated is the human race.
And by extension, biology on earth in this geo-chemical universe.
.
Not doing so good there either, it's amazing how quiet the whole issue of self driving cars has gone since Musk admitted to the US transportation board that his system was only level 2. He's now trying to mislead the public in another way by calling it "Full Driving Experience". It's certainly that if it collides with something!is this pedestrian about to cross the road? etc.
If we are still using tanks, since both modern missile warfare and guerilla warfare so often renders them useless. Seen long ago in the Russian occupation of Afghanistan, again in Basra, Iraq where even civilians stopped our tanks, and most recently in Russia's attack on Ukraine.As you said before SW , a good system will cost several times the rest of the car but is already cheaper than a bus or a big truck. Tesla won't go there, but plenty of others will see the benefits like battle tanks.
We discard, improve, attach or expand theories if they don't match observations - Newton's work ceased to be useful at quantum and cosmic scales... and scale is what makes machine learning work: In the late 90s, I made a simple AI in Perl to select and set astronomy photo of the day as a desktop photo on my laptop, based on my scoring daily the ones it selected. It worked progressively better, after many inputs. Now thanks to people's (often involuntary) very personal use of the internet, even simple self-learning system can become more effective, given datasets many orders of magnitude larger than was possible to collect before.It's like 20th century physics, until Einstein.
Out of all of the applications you've listed, this one interests me most. Hedge fund "AI" R&D labs must be busy...where will the stock market be tomorrow, next week, next year