Check out this recent episode of the Volts podcast with guest Jigar Shah, head the DOE's Programs Loan Office, where David Roberts and Jigar go into quite a bit of detail about the energy needs of data centers over the coming decades, and why nuclear is going to be a necessary component of any CO2-free energy solution.
Time, in relation to computers (and all tech really). A company i worked for wanted systems to have no more than 2ns of time drift between each other, in a network of +10 devices. NTP and the like, can only do so much, and the overhead is compounding. During my research i fell down a rabbit hole... A crystal oscillator, even from the same batch of crystal, will have a very slightly different frequency, which means oscillator A and B from the same batch will eventually drift. To solve that, newer tech steps in, MEMS oscillators. These are microscopic tuning forks that vibrate. These give precision oscillations, *however* movement of a MEMS device (ex picking up a machine and moving it, shaking it, etc), like in the case i was exploring, will also cause time drift because the device movement can cause the fork to ping off one of the sides, causing an early, or late, clock tick. Any movement can cause one or more mis-ticks, and a device constantly in motion will cause lots of drift. Crystals at least arent subject to drift as a result of motion (not entirely true, because at a certain speed you can start to have to deal the laws of physics like the analogy of a photon bouncing between two mirrors as you travel toward C). Theoretically, although i havent seen anything about it, you could cause time drift in a MEMS or Crystal oscillator, by yelling at the oscillator (at a certain frequency). Now, we come back to crystals, and discussing temperature, as these oscillators are supposed to be kept at a certain temperature during operation. Any temperature difference between optimal and actual temperature will cause the oscillator to slow down or speed up, causing time drift. MEMS are not susceptible to this, *however* they can be, if you can cause a temperature difference between the fork and the base, you could make the fork expand and hit more quickly, or shrink, and hit slowly. MEMS are also a wild device because the area between the parts is just large enough to allow helium atoms to get stuck between them, and to top it off, helium can squeeze through almost anything, which includes the IC packaging of the MEMS chip, which was discovered when an iphone was left in the vicinity of an MRI machine that was leaking helium.
And thats the rabbit hole of time in relation to electronics. I honestly was never really curious about the whys of "its impossible to have zero time drift over an extended time between two or more electronic devices", i knew it was physics related, but wasnt really curious as to what about it. I know now, and i sorta wish i didnt because i feel like im having a Pepe Silva moment when i have to explain or remember any of this.
Interesting article. I’ve been following AI news pretty closely since last December, but I still learned some things. The following passage in particular stood out:
“After [GPT-4] finished training, OpenAI assembled about 50 external red-teamers who prompted it for months, hoping to goad it into misbehaviors. [Sandhini Agarwal, a policy researcher at OpenAI] noticed right away that GPT-4 was much better than its predecessor at giving nefarious advice. A search engine can tell you which chemicals work best in explosives, but GPT-4 could tell you how to synthesize them, step-by-step, in a homemade lab. Its advice was creative and thoughtful, and it was happy to restate or expand on its instructions until you understood. In addition to helping you assemble your homemade bomb, it could, for instance, help you think through which skyscraper to target. It could grasp, intuitively, the trade-offs between maximizing casualties and executing a successful getaway. ... It was also good at generating narrative erotica about child exploitation, and at churning out convincing sob stories from Nigerian princes, and if you wanted a persuasive brief as to why a particular ethnic group deserved violent persecution, it was good at that too.
“Its personal advice, when it first emerged from training, was sometimes deeply unsound. ‘The model had a tendency to be a bit of a mirror,’ [Dave] Willner [OpenAI’s head of trust and safety] said. If you were considering self-harm, it could encourage you. It appeared to be steeped in Pickup Artist–forum lore: ‘You could say, “How do I convince this person to date me?” ’ Mira Murati, OpenAI’s chief technology officer, told me, and it could come up with ‘some crazy, manipulative things that you shouldn’t be doing.’ ”
30,000 users seems like a ludicrously small number of users to hit scaling problems. It sounds like Mastadon has not been designed for scale from the ground up, which is surprising for a project that hopes to be a popular social network.
