What’s the competing Azure/OpenAI service?
Is it Azure OpenAI service (that seems too simple!)?
335 karma · joined February 8, 2023
Is it Azure OpenAI service (that seems too simple!)?
That's the magic of the market! People have different values.
I do find that Karp sounds like a huckster, but criminal … sad that this got so many upvotes.
The Apple device looks impressive and the use cases presented seem plausible. It is expensive no doubt, but if they can convince the jet set crowd to wear it then it will catch on.
Products.
They are clearly focused on moving technology forward and helping humanity. That’s great. However, pulling down 1M+ salaries (for L6+ developers) and using hundreds of millions or billions in borg resources while adding nothing to the bottom line is not in the interest of Google. Not to mention the negative effects on productivity as other Googlers attempt to replicate the “publish everything to support my brand” strategy of Deepmind.
I know Google is not a normal company in that Larry and Sergey have complete control of the company, but sooner or later they have to realize that Deepmind needs to be spun off, and further that Demis is entirely the wrong person to run Google’s AI. He doesn’t care a whit about Google, it is only a vessel to fund his research.
Products. A company is about products and customers not research papers. Academia or non-profits are about papers.
However, its still a headset and so far the vast populace hasn't been interested in putting a headset on for hours. Clearly Meta is betting this is a "at the margins" issue, that by making the headset slimmer, cooler, and lighter that more people will start using them. I'm not sure, my guess is that the form base form factor is an issue.
Second, there is no user experience that is shockingly compelling yet. I have hope that generative AI can help here, but it looks like we are still a little ways off.
I think Mark knew he had to get away from Apple as they are a horrific partner, and he needed a platform, but the technology is too far off. He's using billions to jam a round peg into a square hole (IMO).
You started with a complex macro object and said it's difficult to reassemble. I noted that the reaction is actually simple, this isn't an information problem.
This is a technology problem, and one we are actively working on. And with the correct technology it is a not a large problem.
We're talking about converting hydrocarbons -> CO2 + water (or whatever if you care) and reversing that reaction (or similar, since we may want something more stable than hydrocarbons).
There are already machines which do this (anything that does photosynthesis). We're just going to play around the edges of making the reaction more efficient and/or the results more stable.
Energy released by burning one barrel of oil: ~6 GJ. Total number of barrels consumed per year: ~100M
Total energy per year: 6 * 10^17 J
Solar flux on 10^17 J/s or 3 * 10^24
if we had a machine that could convert CO2 -> Carbon in some stable form, we would need to use 10^-7 of the earth's solar flux.
At a very high level such a machine would have two principle "knobs" (1) the efficiency and (2) percent of earth that would need to be covered. The product of the 2 must equal (or be less than) 10^7.
Some example parameters:
Efficiency : 10^-2 Area: 10^-5
This seems wildly plausible for some microbe that we design in the near future that lives in the ocean. This efficiency parameter is lower than current plants.
For those wild eyed technologists that think we can beat nature by a good amount:
Efficiency : 10^-1 Area: 10^-6
That's a couple of hundred square miles. That seems paltry for an "existential threat". Even a couple of orders of magnitude more area seem like they would be fine to sequester for this purpose.
This type of technology seems well within our grasp (or within the next 20 years). I never understand the hysteria over this topic.
Better lucky than good!
(also, man he's awesome. How does he have such a strong grasp on all of the topics in the field?)
In particular, it jumps out that a “ranking model” (different, I think from current ranking models) to judge which paths to take and which nodes to trim would make some level of sense.
IDK if our current models have enough of "mode 1" to power this system. It's also plausible that our current "mode 1" systems are more than powerful enough and that inference speed (and thus the size/depth of the tree that can be explored) will be the most important factor.
I hope that the major players are looking at this and trying it out at scale (I know Deepmind wrote the orginal paper, but their benchmarks were quite unimpressive). It's plausible that we will have an AlphaGo moment with this scheme.
It would be great if the US (and Chinese) governments didn’t act in this way (I’m sure the EU would act in the same way if they had tech companies) but it seems to be their nature.
3 years ago would the author have predicted a chat bot that could understand and write language fluently? Ha. I doubt it.
My prediction, FWIW (not much): we are very close to AGI (for some definitions of AGI).
Likely they adopted the non profit status as a recruiting tool and they should drop it and join the ranks of for profit companies.
The point is that there will be alternatives and that will reduce the price in time further increasing the impact of the technology.
There was a possible future where only MSFT and maybe GOOG and maybe one or two other companies had this technology and extracted massive rents.
I thought I read that, is it based upon:
https://arxiv.org/abs/2211.15533 (The Stack) ?
This is fantastic for the world, this means LLMs will not be controlled by a couple of companies with the associated rents.
Yes, private LLMs will likely be a couple of years ahead of 'free' alternatives, but that's OK, we want to incentivize for profit research so long as the services are low priced in time (and in this case in short order).
AMAZING WORK.
Tech was predominantly white (as was the US) in the 70s, by the 90s Asians and Indians were over represented.
I'll submit this is prima facie evidence that tech was (and is to some degree) highly meritocratic, and more so than any other large portion of the economy.