Jump to content
NotebookTalk

Etern4l

Member
 View Profile  See their activity

  • Posts

    1,896

  • Joined

  • Days Won

    13

 Content Type 

Everything posted by Etern4l

  1. Etern4l
    UK watchdog to review AI models like ChatGPT - FT The same regulator just blocked Microsoft's takeover of Activision-Blizzard. Promising.
  2. Etern4l
    If this is for air cooling then they look a bit sub-optimal from the performance perspective. Need to have some features which increase the surface area, like grills or something. I would understand the simple block design with holes for WC.
  3. Etern4l
    I mean it's not hard to see where things are going. The simple question to answer about the very short term impact is: what percentage of the given job could the current, or a slightly more advanced AI do much faster and at a close to zero cost. The expected job losses in that area roughly correspond to that fraction. Remember, that - as per the Industrial Revolution - the overall global economy will shrink as a result (because of the extra unemployment), so the affected will be unlikely to find alternative employment (on average) in any but the least desirable areas. The process has already started, although it's just the tip of the tip of the iceberg. You are absolutely right, what people are failing to grasp is the size of the snowball, which will grab and smash pretty much the entire workforce: from scientists to plumbers and farm workers (just google agricultural automation if in doubt).
  4. Etern4l
    That's all well and fun, but what if creating those daunting Linux scripts was a significant part of the job and a part of the value proposition. If that's effectively automated, what is the added value left? Consider a larger company which employs 10, 100, or 1000 people in a similar role. If let's say their productivity improved by 50% (conservatively), they will reasonably have to consider axing 5, 50 or 500 people, unless there is something else for them to do that AI can't. Moving around might not be easy, as there will be people in other areas in a similar position. I very much doubt the US will be lagging, quite the contrary - it will be at the forefront of the change, being the barebone free market system dominated by large corporations. See the update from IBM above, and consider the layoffs situation in the US tech sector which happened to coincide with the reveal of ChatGPT. AI is often compared to the Industrial Revolution. The interesting thing there is that industrialization initially resulted in a loss of jobs and economic output. Eventually, people managed to use their brains and re-adjust. This situation is fundamentally very different. There won't be anywhere to run, but to lower skill and pay physical work, until humanoid robots arrive that is. Clearly you haven't had a chance to go through all the content and videos, but let me repost a brilliant quote from the Max Tegmark (who unfortunately makes the mistake of trivialising the economic impact, a common theme with the self-interested AI community) interview on page 3: "We're rushing towards a cliff, but the closer we get, the more scenic the views are."
  5. Etern4l
    Thanks. Updated the list The system was more watertight on NBR. Possibly the only technical advantage of the old site. Not a biggie, but worth addressing if possible. Alternatively, perhaps firmer enforcement of a reasonable standard of posting might be worth considering, so that even fairly patient users are not forced to use the feature for sanity's sake. Without this, the noise is still there for any new and uninitiated visitors to see, could be off-putting.
  6. Etern4l
    I agree so much, I bothered to create a thread on this. Shamelessly plugging.
  7. Etern4l
    I am calling BS in their policy, just wanted to entice people to interact with the platform. They have no real incentive to delete real user accounts. In theory they should be identifying and deleting multi-accounts and bots, but again, not clear they want to do that. Content is content, and free speech for bots too (might want to stay on their good side before Starship is ready for the ultimately hopeless evacuation of the select few to Mars). BTW I'm abandoning the use of "likes", upon the somehow only recent realisation that it's one of the key tools social media use to promote unhealthy user dependence on the platform.
  8. Etern4l
    One AI reaches that level of capability (luckily it's not remotely there at the moment, it can "only" solve much smaller and simpler problems), we are done as species. We have 5-20 years to save ourselves from being deprecated - according to Geoffrey Hinton, one of the fathers of modern AI. Sounds like a reasonable timeframe to me, although I would bet on the shorter end just to be safe. No doubt we are quickly approaching the most precarious time in our history.
  9. Etern4l
    I mean it would be good for someone to make a AAA remake of this, however, failing that if people want to take it to powerful AIs they can do so in real life these days (the primary targets being Microsoft and Google at the moment).
  10. Etern4l
    I think the laptop stuff is consistently lagging behind desktops, where we saw similar issues around 12th gen launch, i.e. 3600-4400 CL38 speeds no matter what modules. Now (several BIOS updates later) running at 5200 CL30 on old Hynix dies and a Z690 board. I believe also on the 12th gen but not 100% sure, but 4800 is def solid.
  11. Etern4l
    Hello, Would it be possible to close gaps in the ignored users function (in the suggested order of importance): * Reactions from the ignored users shouldn't be visible * Noise from ignored users is still unnecessarily shown in the activity feed * Filter out from Last post thumbnails (a couple of users showing on the majority of threads on the site especially in off-topic) * Ideally we would also be able to filter out threads started by the ignored users Thanks guys. @Hiew
  12. Etern4l
    Really not friends with NPR.
  13. Etern4l
    Right, all against the backdrop of inflation (the 99% getting poorer, resulting in less disposable income) and consumer PC sales slowing across the board.
  14. Etern4l
    Here we go: IBM to pause hiring in plan to replace 7,800 jobs with AI - Bloomberg News Presumably pretty decent jobs - poof! ChatGPT warning knocks $1 billion off market caps of education giants Pearson and Chegg Translation: the rich getting richer by shorting the education sector and thus making an informed bet that AI will take it out.
  15. Etern4l
    Star Citizen is basically a scam, no?
  16. Etern4l
    Like i said, it's quite complicated. Here is a single-item reading list google may have failed to suggest to you for some reason: https://www.scribd.com/document/452452273/Semiconductor-Reliability-handbook-Renesas-pdf Signing out of this.
  17. Etern4l
    LOL. Alright, I give up :) To end of a friendly note, let's just say it's not as simple.
  18. Etern4l
    Welcome to non-linearity in physics. When I say risk, I mean risk over typical lifetime. It could be that you are reducing MTBF from 100 to 50 years, you still don't care.
  19. Etern4l
    It's not binary. Overclocking may increase the marginal risk of a fault occurring, depending on how much and what kind of additional stress is generated, and how long it is applied for. In some cases, the risk is close to zero, in some cases, it can be significant (e.g. shunt modding a 600W card to 1kW, and running this 24/7 lol)
  20. Etern4l
    All I'm saying is that ChatGPT is not the underyling model, the underlying model has no human/programmer rules imposed on it.
  21. Etern4l
    That's nor completely true. In fact the raw model underlying GPT actually expresses "balanced" opinions based on the vast data it was trained on, which is why it also underwent additional "alignment" training and thus dumbed down by OpenAI to be politically correct, avoid certain topics etc. Things are already bad in some sense for example: in the social media, criminals can copy your voice etc. The point is to ensure things don't get even worse in the near future (rather than ways off).
  22. Etern4l
    Yep, it's not super-common but still used in the UK, mainly England, Scotland and NI. Not sure about Wales.
  23. Etern4l
    Warning of AI’s danger, pioneer Geoffrey Hinton quits Google to speak freely The guy is a very well known academic in the field. Good data point on what's going on at Google, and whether the dangers of AI are real.
  24. Etern4l
    Yikes, that makes no sense. Would expect a performance penalty when using 4 SODIMMs, but not a general penalty on 4-DIMM systems. Are we saying that Lenovo is falsely advertising 4800 regardless of memory capacity? Anyway, even if the P16 performs as advertised, 4800 with 2 SODIMMs is not great. GT77 is 13th gen so perhaps the new IMC is making some difference, although 12th gen would go up to something like 6600 on some desktops.
  25. Etern4l
    Yep, I've heard about a concrete case from a family member already, about a month ago. "Distressed daughter" call. That panicky call from a relative? It could be a thief using a voice clone, FTC warns Mum, help: Nina made three bank transfers before realising she had been scammed It's fascinating, and obviously unsettling, to see parts of sci movies step by step becoming reality before our very eyes (a complete shapeshifting T1000, the robot, rather than the Nvidia video card, is probably some time away, fortunately!).
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Terms of Use