HNHacker News
TopNewBestAskShowJobs

skjoldr

212 karma · joined December 3, 2019

submissionscomments
skjoldr··on LaMDA is not sentient
Yes, it is partially about whether human models of reality (encoded in the relationship between word tokens) can approximate reality itself to such a degree where the quantity of ingested information can substitute for its quality.

Human brains and senses are obviously qualitatively different, but AIs have the advantage of unimaginably massive bandwidth of incoming textual information.

Now, one of the hardest difficulties here is setting up the goal structure for this kind of AI. Just filling in the blanks is obviously insufficient.

skjoldr··on LaMDA is not sentient
What I really wonder is what will happen if we feed an AI similar to LaMDA or GPT-3 a bunch of books on programming and system administration, re-written in a certain way that is better suited for AI processing, and its own code. Would it be able to very primitively reason about itself?
skjoldr··on LaMDA is not sentient
The only real difference between what you are describing and GPT-3 is that your dataset from which you draw all your opinions about the world is unique and comes from the sensory input you had experienced in your life, stored directly in memory. Whereas GPT-3's dataset comes from slices of many other peoples' sensory experiences, stored in language. If it was possible for a single human to produce enough words that describe their own experience to rival the volume of GPT-3's dataset, then it would be possible to train a model that would essentially "hold" that person's opinions. But I don't think such consistency is required for intelligence. It is simply arguing that an intelligent being must only experience the world at a single point in space, at a time, while this kind of AI takes in the data about experiences that happened in many different points in space at many different times. It seems kinda arbitrary.

What if AR tech develops to the point where you can experience "being" in other point in space through artificial sensors (say, on a robotic or drone chassis), while also being able to see and hear your real surroundings to some degree? Then you will be able to experience the same real event from multiple vantage points. Will you be able to form two different opinions on what "really" happened? Which one would be the "correct" one?

skjoldr··on LaMDA is not sentient
Your argument, then, hinges on the fact that a human being can only be in one place at a certain time, thus their experience of the world is unique. An agent that takes in experiences of thousands of people at thousands of places at different times lacks this quality. Naturally such an agent will not have a single opinion about the world. But, is it not better to be able to consider many different viewpoints at the same time? Why do we have freedom of speech and expression, why do we learn other languages, if not to be able to hear other people's points of view about the world, even though we may not have had the particular unique experiences that shaped them? Isn't being able to take a broader perspective and being able to argue for any side of a debate regardless of personal preferences, a sign of greater intelligence in humans? Then why an AI that thinks entirely in this way should not be considered at least somewhat intelligent?
skjoldr··on LaMDA is not sentient
Why do you think that combining small slices of thousands of other people's intelligences does not produce its own form of intelligence?

When you ask an AI "do you like what I am wearing?", knowing what an AI is, the only logical explanation for me would be that you, in fact, want to know the average opinion of other people about your dress, since that is the only answer an AI can give. Which is a perfectly good question in and of itself that we also ask other people. In that case, a good quality answer would entail a past history of communication with many other human beings about their tastes in clothing, producing the ability to accurately guesstimate if other people will like what you wear. But this is exactly the realm in which this kind of AI can surpass human beings simply because it can "talk" to many more people from many more locations across time, thus producing a better answer for you. Is this ability to collect and process information about other people not a type of intelligence? There are many other types, yes, but it is possible to construct a system that consists of more than one type of neural network that can relay data among themselves.

skjoldr··on LaMDA is not sentient
Human languages exist to compress the world around us in a way that will survive being sent over sound waves and uncompressed in another brain that experiences a similar enough world around them.

Meaning that with enough words and their relations, sufficient complexity forms for its processing to result in intelligence.

Remember that an AI can process far more textual data than we can. It is still orders of magnitude smaller the amounts of information we process through our senses, but it is still a start.

skjoldr··on Zulip – Threaded real-time chat for distributed teams
It did require a lot of effort for Element, Element is a competitor to Zulip, and I found the UX around E2EE of Element to be perfectly acceptable and consistent with UX of the previous versions, so your comment was also a bit disingenuous IMO.

It is obvious that E2EE would involve a big rework of the codebase, but it is neither impossible nor unnecessary. If there are architectural problems with it, I have to question the original developer intent, as a lot of stuff that E2EE requires to work also improve privacy in general.

skjoldr··on Zulip – Threaded real-time chat for distributed teams
If I want my messages to stay away from the public Internet, why would I even want to receive message content over unencrypted email? That makes no sense. Receiving notifications while I am not even logged in also makes little sense because I obviously did not consent to them. And having optional E2EE lets you choose which kinds of messages you don't mind leaking.
skjoldr··on Zulip – Threaded real-time chat for distributed teams
Wait a second, Element has had working optional private and group E2EE for a while, why can't Zulip?
skjoldr··on World's only An-225 destroyed by Russian airstrikes
You know nothing about the extent to which it was or was not taken apart or whether it was cleared to fly. Stop.
skjoldr··on Show HN: MdSilo – A knowledge silo that runs in your web browser
"Note: by default, Notabase has a cap on the number of notes you can create. You can circumvent this by inserting an active subscription for your user id in the subscriptions table."

Not really "self-hosted", then.

skjoldr··on Risk of myocarditis following sequential COVID-19 vaccinations by age and sex [pdf]
No. The messaging around Omicron was explicitly fear-based. Now even the media is forced to concede that it is mild.
skjoldr··on Ask HN: My client want an agent on my laptop. Is this the new normal?
If you can afford to spend a bit of money on the problem, it's possible to use something like PiKVM or KVM-over-IP to just leave a stack of client laptops or mini-PCs out of the way somewhere and connect to them remotely in a reliable way, so you can reset the machine if the remote desktop software fails.
skjoldr··on Germany locks down unvaccinated people, as leaders plan to make shots compulsory
These stats do not show the difference between taking a shot but still being infected within two weeks after it, and not taking any shots at all and being infected. If you're going to blame the second group for filling up ICUs, you cannot lump them together.
skjoldr··on New emails released in the McDonald’s ice cream machine lawsuit
It is the franchisee who pays for repairs. Of course it was McD's idea. They make a cut.
skjoldr··on PiBox Mini – Modular Raspberry Pi Storage Server
Pi has limited PCIe lanes to devote to additional controllers. PiBox lacks USB3 for this reason -- all the throughput is already filled by SATA3. 2.5Gbit is likely to be mutually exclusive to either.
skjoldr··on PiBox Mini – Modular Raspberry Pi Storage Server
It's weird that there is no tier that offers everything from the Standard tier (case, fan, display, power supply etc.) minus CM4. I'd rather trade Wi-Fi capability for more eMMC storage, considering this thing will likely sit right next to my router anyway.
skjoldr··on Canon sued for disabling scanner when printers run out of ink
I was recommended this ink by local printer servicemen. It turns out they really don't like working on broken inkjets (it's really messy and it's not guaranteed that a dirty nozzle will work even after ultrasonic cleaning), so they tend to give sound advice on how best to keep them running, including ink choice. It really depends on the printer manufacturer and model so YMMV.
skjoldr··on Canon sued for disabling scanner when printers run out of ink
I've got two A3 inkjets, an older Epson Stylus Photo 1410 with 3rd party ink tanks installed and externalized waste ink container mod, and a newer Epson L1800 with built-in ink tanks, at my workplace. Both are 6 color photo printers. I've kept them running for years with few problems. Long-term maintenance consists in keeping the nozzle clear of dried pigment, chaniging the rubber paper feed roller every couple of years, and keeping an eye on the ink levels in the tanks and the waste ink container. The 1410 had printed over 110,000 pages by now and is stil going strong. Printing with 3rd party inks is incredibly cheap, I use Korean Ink-mate ink with these two and it costs a fraction of a cent per page.

Now, it is true that inkjets are fussy and you need to keep printing with them to keep them operational, but I managed to clean out a very dry nozzle twice with nothing more than a pack of lint-free paper, some specialized cleaning fluids, a pack of syringes, and a thin rubber tube, simply using guides found on the Internet. And I haven't had any mechanical failures yet that weren't caused by the paper feed roller being worn out (which is expected) despite both printers being used daily for years.

Maybe I'm just lucky, or maybe Epson did really good with these two models in particular, I don't know.

skjoldr··on Canon sued for disabling scanner when printers run out of ink
It is rather easy to keep an inkjet printer from gunking up if you buy a set of refillable cartridges, fill them with a 50/50 mix of demineralized water and inkjet cleaning fluid, install them and run a cleaning cycle after you finished a printing session. No ink in the nozzle means no gunk. This trick also works well for cleaning nozzles, since the cleaning fluid is a surfactant and it breaks up clumps of dried pigment better than just a stream of ink.
skjoldr··on A good livestreaming setup is a well-tailored suit
You invest your time into your appearance so other people won't need to waste their time trying to determine if you're trustworthy. Appearance is a heuristic. Raging about it is counter-productive.
skjoldr··on Linus Torvalds on mRNA Vaccines
Because a passport represents a centralized storage of medical data, and that's always a bad idea.
skjoldr··on Linus Torvalds on mRNA Vaccines
Yet people like this who just don't want to inject anything that hasn't been studied for at least a couple of years, don't coach our opinions in anti-vaxx language. Stuff like "new species" is ridiculous and masks the actual rise of surveillance totalitarianism which uses Covid as an excuse.
skjoldr··on Reclaim Windows10
Is there any use for SmartScreen when running a proper antivirus, as you should?
skjoldr··on Ryanair’s statement on diversion to Minsk Airport
I think any reasonable person understands that international politics are inherently driven by national interests and thus are hypocritical by default, because often there are no perfect solutions that are acceptable to everyone. Least you can do is to call hypocrisy out when it happens, and ask yourself whether it was at least done in your interest as a citizen or not. If you want to compare political systems, do it systematically, not on case-by-case basis. Russia had jailed many more political activists and has far more draconian laws against freedom of speech and freedom of assembly than Western countries, even if Western countries sometimes unjustifiably limit these freedoms as well.
skjoldr··on Ryanair’s statement on diversion to Minsk Airport
Human rights violations are human rights violations, regardless of who commits them. In the Russian space, they occur daily, in the West they are rarer. What is the point?
skjoldr··on Tesla is found guilty of throttling charging speed
There is info that shows phone Li-Po batteries last much longer if they are always kept at 20-80% charge at least, and that both the default charging settings and the marketing materials advertising battery capacity are biased towards bigger numbers. I'm not sure what is the issue here, then, beyond bad communication, especially if only the older battery packs are affected. It is to be expected that if the customer wants to keep using the pack for a while longer, they need to limit how much they cycle the battery every charge. A better solution would be to advertise the average longevity of the Tesla battery packs along with their mileage, but I'm not sure if their target audience really cares about that.
skjoldr··on Ryanair’s statement on diversion to Minsk Airport
Yes, Snowden's treatment is appalling. What does it have to do with this case?
skjoldr··on Hetzner has banned crypto mining and Chia farming
Mining is not general purpose.
skjoldr··on Hetzner has banned crypto mining and Chia farming
You do get abuse notices that you have to respond to before your server gets isolated from the Internet. If you don't, you can request unblocking later from support staff.
← PreviousPage 4 of 5Next →