You are making a wrong leap from non-deterministic process to uncontrollable result. Most of the parallel algorithms are non-deterministic. There might be no guarantee about the order of calculation or even sometimes the final absolute result. However, even when producing different final results, the algorithm can still guarantee characteristics about the result.
The hard problem then is not to eliminate non-deterministic behavior, but find a way to control it so that it produces what you want.
Life and a lot in our universe is non-deterministic. Some people assume science and mathematics are some universal truths rather than imperfect agreed upon understandings. Similarly many assume humans can be controlled through laws, penalties, prisons, propaganda, coercion, etc. But terrible things happen. Yes, if you set up the gutter-rails in your bowling lane, you can control the bowling ball unless it is thrown over those rails or in a completely different direction, but those rails are wide with LLMs by default, and the system instructions provided it aren’t rules, they are an inherently faulty way to coerce a non-deterministic system. But, yes, if there’s absolutely no way to do something, and you’re aware of every possible way a response or tool could affect things, and you have taken every possible precaution, you can make it behave. That’s not how people are using it though, and we cannot control our tendency to trust that which seems trustworthy even if we are told these things.
No, Science is a means of searching for those truths - definitely not some 'agreed upon understanding'. It's backed up by experimentation and reproducible proofs. You also make a huge bogus leap from science to humanities.
Scientific method is the process. Science itself includes the study and compendium of understandings, based on a belief system that includes shared understandings just like mathematics. The foundation of these are philosophical beliefs that we can know and understand these things. For example, on a metaphysical level, if the world around us were a simulation, then science could provide understandings about that simulated universe, but not about that which is simulating it.
This I'm afraid is rubbish. Scientific proofs categorically don't depend on philosophical beliefs. Reality is measurable and the properties measured don't care about philosophy.
But those are still approximations to the actual underlying reality. Because the other option (and yes, it's a dichotomy) is that we already defined and understood every detail of the physics that applies to our universe.
Being exact doesn't mean it is not an approximation, which was the initial topic. Being exact in science means that 2+2=4 and that can be demonstrated following a logical chain.
But that doesn't make our knowledge of the universe exact. It is still an approximation. What it can be "exact" is how we obtain and reproduce the current knowledge we have of it.
To our current knowledge, no. But maybe we are missing something, we cannot know. Did infrared light or ultrasound start to exist only when we realized there are things our senses cannot feel?
Well, sometimes Microsoft decides to change your settings back. This has happened to me very frequently after installing Windows updates. I remember finding myself turning the same settings off time and again.
The "fuck you, user!" behavior of software companies now means there's no more "No", only "Maybe later". Every time I update Google Photos, it shows me the screen that "Photos backups are not turned on! Turn on now?" (because they want to upsell their paid storage space option).
It also now bugs me to do face scanning every so often too
And unlike most things, both prompts require you to explicitly click some sort of "no", not just click away to dismiss. The backup one is particularly obnoxious because you have to flip a shitty little slider as the only button is "continue". Fuck. Off.
Silicon Valley companies are like a creepy guy in the nightclub going up to each woman and asking "Want to dance? [Yes] or [Ask Me Again]". The desperation is pathetic.
I mean at this point I think it's really just utter incompetence over at Microsoft to design a system that can be updated without breaking it. They have never actually cared about solving that problem.
If they had taste, someone opinionated over there would knock heads before shipping another version of windows that requires restarts or mutates user settings.
A joke in the Windows 95 days was "You plugged in a mouse. Please restart your computer.". A few weeks ago I plugged in a Logitech wireless mouse receiver, Windows 10 installed the drivers automatically, and finished with "To complete the installation of the software, please restart your computer"...
There are two kinds of consultants: those who write code and those who only give advise. It seems to me that those who only advise lost their market to LLMs pretty completely.
I'm surprised reading that the Iranian's regime concerns are centered on WhatsApp sharing information with Israel. It is much more likely that WhatsApp have 0-day vulnerabilities used by the Mossad to gain the info than WhatsApp actively sharing it.
> Iran banned WhatsApp and Google Play in 2022 during mass protests against the government
So more than fearing Israel, they actually fear the public that has an encrypted communication channel that can't be tapped by their police. Explains a lot.
Russian soldiers participating in SMO reported multiple times that after they exchange texts and photos with their relatives in WhatsApp, that information ends up in Ukrainian military HQs next day. With photos. And later used, for example, to harass the relatives of the soldier.
Could be some other mechanism (e.g. Google Drive or some other kind of malware), hard to be sure in the world, where since 2011 Snowden's revelations, bugs are placed my NSA and CIA everywhere, starting from hardware and firmware.
Russian soldiers chatting with their family know what precisely happens in the Ukrainian military HQ the next day? That sounds too crazy to be even remotely true, and too convenient a story ("they are harassing your family, go murder someone").
After arrest of Durov in France I think the same. Doesn't change the fact that any single Russian soldier is not using WhatsApp. It's considered compromised with very convincing evidence, along with some other apps (I don't remember all the names, possible to find in military news channels if you know Russian).
There were evidence of leaks of soldiers' messages from WhatsApp.
I do not know for sure if there are leaks from Telegram, but I think that Durov was pressed for enabling backdoors for CIA when he was arrested in France.
It's not like there's a registry of compromised apps in Russia that is available to public. When there were WhatsApp leaks it was widely published on Russian news channels, that's why I know about it and tell everyone to not trust WhatsApp.
But I don't closely follow every news item from the front, I know that some other apps are also considered compromised (mainly some navigation apps), but I don't know the whole list or if Telegram is being used on the front.
I guess the safest option is to not use telephone at all at war, since there are probably backdoors everywhere starting from cell modem firmware. I guess that since about middle-end of 2022 all military personnel has the same opinion and stopped using smartphones altogether or use them in flight mode with communications disabled.
> but I think that Durov was pressed for enabling backdoors for CIA when he was arrested in France.
Durov was pressed from every side from the beginning, no doubts about this. But CIA or FSB cannot demand it legally even with secret court orders. There were news that since arrest in France he started providing some info on criminals. Not sure how far did he go. But it's good to know that Whatsup is fully cooperating with governments even when they don't have to. Putting people at risk.
As for smartphones, there are so many security holes that it's impossible to secure. Many harmless applications are calling home, pictures are geo-targeted, cell towers can request info from connected phones, every update can turn OS or app into trojan. If infected it's a perfect audio and video collection device. Put together with soldiers' IDs it shows where the unit is. Add to that CIA and alike working with weakening everyone's security and getting priority access to sensitive information.
Not really in the space we're discussing, i.e. usian and iranian everyday phone applications. Small private actors aren't going to pop the TLS on messages anyway, and the states involved have or can compromise the system as such.
I am not involved in scraping, but to me this sounds like simply another tool in the arsenal. They say it's hard for the scraper to realize it has been caught this way because it's not being blocked. However, I don't see anything preventing scrapers from implementing heuristics to realize that.
Detecting the actual AI generated content is not an easy problem. Not following deep links and recognizing the particular website template and structure is easier. I really feel a monoculture of anti-bot tools can defeat their effectiveness. When you have to optimize for Anubis, Nepenthes, Quixotic, and Cloudflare, each independently evolving and different in method and implementation, it might just be practical to give up.
This means the leading UI for LLMs - the chat - is the wrong UI, at least for some of the tasks. We should instead have a single query text field, like in search engines, that you continue to edit and refine, just like in complex search queries.
I like zed's approach, where the whole discussion is a plain text file you can edit as any text, which gives you the ability to change anything in the "discussion" regardless if it was generated by you or the llm. It makes stuff like that much simpler, ie you can correct simple stuff in the llm's response without unecessary back and forths, you can just cut parts out of the discussion to reduce context size or guide the discussion where you actually want removing distractions etc. I don't understand why the dominant approach is an actual, realistic chat interface where you can only add a new response, or in best case create "threads".
> I don't understand why the dominant approach is an actual, realistic chat interface where you can only add a new response, or in best case create "threads".
I'm not 100% sure either, I think it might just be a first-iteration UX that is generally useful, but not specifically useful for use cases like coding.
To kind of work around this, I generally keep my prompts as .md files on disk, treat them like templates where I have variables like $SRC that gets replaced with the actual code when I "compile" them. So I write a prompt, paste it into ChatGPT, notice something is wrong, edit my template on disk then paste it into a new conversation. Iterate until it works. I ended up putting the CLI I use for this here, in case others wanna try the same approach: https://github.com/victorb/prompta
I wonder if people in general would have a healthier understanding of LLMs if this mode of interaction was more common. Perhaps it would be more clear that the LLM is a very souped up autocomplete, instead of some other mind to debate.
Yes I think the same. It really felt demystifying the experience for me when I first tried zed, and I already had that belief about LLMs. But when you use an LLM through a normal chat interface, it feels different, and they are also tuned to feel that way, like talking to somebody. Maybe this is why this approach is avoided by those who have some stake on AI, even if it is better UX.
I've found the most useful LLM UIs for me are tree-like with lots of branches where you go back and forth between your prompts. You branch off anywhere, edit top or leafs as you go.
If one branch doesn't work out you go back to the last node that gave good results or the top and create another branch with a different prompt from.
Or if you want to ask something in a different direction but don't want all the baggage from recent nodes.
I still think there is value in chats and retaining context. But there is also value in starting clean when necessary. Giving users control and teaching people how to use it is the way IMO.
The problem with retaining context is that it gets polluted. That pollution gets you into a latent space with errors, which probably not where you want your next token prediction to be sourced.
The reasonable alternative is a chat interface that lets you edit any text, the AI response or your prompts, and regenerate from any point. This is why I use the API "playground" interfaces or something like LibreChat. Deepseek at least has prompt editing/regeneration.
> This means the leading UI for LLMs - the chat - is the wrong UI
For coding, I'd agree. But seemingly people use LLMs for more than that, but I don't have any experience myself. But I agree with the idea that we haven't found the right UX for programming with LLMs yet. I'm getting even worse results with Aider, Cursor and all of those, than just my approach outlined above, so that doesn't seem like the right way either.
And yet, it's the best we currently have. I donate to them. We can come with demands of how it should be managed, but it should not prevent us from helping them.
If you poke around at what US government agencies are doing, and what European countries and non-profits are doing, or even do a deep dive into what your local library offers, you may find they no longer lead the pack.
They didn't even ask for donations until they accidentally set fire to their building annex. People offered to help (SF was apparently booming that year) and of course they promptly cranked out the necessary PHP to accept donations.
Now it's become part of the mythology. But throwing petty cash at a plane in a death spiral doesn't change gravity. They need to rehabilitate their reputation and partner with organizations who can help them achieve their mission over the long term. I personally think they need to focus on archival, legal long-term preservation and archival, before sticking their neck out any further. If this means no more Frogger in the browser, so be it.
I certainly don't begrudge anyone who donates, but asking for $17 on the same page as copyrighted game ROMs and glitchy scans of comic books isn't a long-term strategy.
I simply cannot give up the option of zapping distractions off of my screen. I really cannot understand how people can use Youtube or even a Youtube embedded video without zapping away the distractions. There is no way I'm coming back to Chrome if they don't support manifest V2. It's Firefox for me.
I'm also using it to remove distractions from stack exchange sites like hot network questions. I can stay focused on solving my problem much better if my eyes can't wander to interesting unrelated stuff.
Many people are annoyed but 1) don't know there is any other way, 2) got used to it and the hassle of it is still worth it for them to get to watch the stuff they want.
The parent comment is talking about distractions, not ads. YouTube has plenty of those, even embedded YouTube videos, unless you pause the video before it ends. uBlock Origin Lite cannot block elements except through packaged rulesets, and while there are some ad-blocker lists that are meant to block annoyances on pages in addition to ads, everybody has a different idea on what is an annoyance on a webpage.
exactly. I was not referring to ads, but to annoying suggestions in embedded videos (also when the video is paused!), and even to the long and mostly useless suggestions list on the right of the screen. I want to use YouTube as a useful tool, not to waste my time in endless loops of "oh, that looks interesting!"
If there was a similar product that does not upload any of your extremely personal data, like whether you're now in your bed, to some server on the internet, would you prefer it?
Sure, there are lots of ways it can be improved. I'd like it to be cheaper too. I'd be happy to switch to an alternative that is just as good but without the Internet nonsense, but SleepMe isn't it. I've got my eyes open for viable competitors for the next time I need to outfit a mattress or when this one dies. For now, Eight Sleep is the best one I've found.
The hard problem then is not to eliminate non-deterministic behavior, but find a way to control it so that it produces what you want.