cwagner

@[email protected]

Profil ze zdalnego serwera może być niekompletny. Zobacz więcej na oryginalnej instancji.

cwagner,

I was confused about that as his Wikipedia page didn’t show anything that bad, but didn’t want to get into that :D

cwagner,

Eh, not sure I agree. Seems to also have been between too little and too much AI safety, and I strongly feel like there’s already too much AI safety.

cwagner,

Using it and getting told that you need to ask the Fish for consent before using it as a flesh light.

And that is with a system prompt full of telling the bot that it’s all fantasy.

edit: And “legal” is not relevant when talking about what OpenAI specifically does for AI safety for their models.

cwagner,

I don’t really care, but I find it highly entertaining :D It’s like trash TV for technology fans (and as text, which makes it even better) :D

cwagner,

Nope

Best results so far were with a pie where it just warned about possibly burning yourself.

cwagner,

No, it’s “the user is able to control what the AI does”, the fish is just a very clear and easy example of that. And the big corporations are all moving away from user control, there was even a big article about how I think the MS AI was broken because… you could circumvent the built-in guardrails. Maybe you and the others here want to live in an Apple walled garden corporate controlled world of AI. I don’t.

Edit: Maybe this is not clear for everyone, but if you think a bit further, imagine you have an AI in your RPG, like Tyranny, where you play a bad guy. You can’t use the AI for anything slavery related, because Slavery bad, mmkay? And AI safety says there’s no such thing as fantasy.

cwagner,

AI safety is currently, in all articles I read, used as “guard rails that heavily limit what the AI can do, no matter what kind of system prompt you use”. What are you thinking of?

cwagner,

If it helps even more: The AI in question is a 46 cm long, 300 g heavy, blue, plushie penis named after Australia’s “biggest walking dick” Scott Morrison: Scomo, and active in an Aussie cooking stream.

cwagner,

Is there anything new in this post that I’m missing?

cwagner,

It’s a substack post. At this point, my quality expectation is

  1. Wordpress - Probably someone who really cares about what they write about
  2. Substack - Either low effort spam like this that gets upvoted for some reason or someone pushing their agenda, hard
  3. Medium - Either spam, wrong, dumb, or too simple. Literally never worth reading.
cwagner,

Heh:

The OpenAI tussle is between the faction who think Skynet will kill them if they build it, and the faction who think Roko’s Basilisk will torture them if they don’t build it hard enough.

– mastodon.social/@jef/111443214445962022

cwagner,

I’d say this is an amazing result for MS. Not only is their investment mostly Azure credits, so OpenAI is dependent on MS, now they also got Altman and his followers for themselves for more research.

cwagner,

Doesn’t really work when none of this was initiated by MS

cwagner,

I don’t mind so much what they did with firing him, but how they did it, and everything since. It just seems extremely unprofessional and disorganized.

cwagner,

They believed that the AI safety work they had done was insufficient.

Considering that every new model seems to be getting worse for anything but highly sanitized corporate usage, I’m not sure that I want more AI safety …

For my usage, I use Chat GPT 3.5 turbo with the march checkpoint because I can’t get the current one to stop moralizing about bullshit instead of doing what it’s supposed to (I run two twitch bots with it). GPT4 used to be okay there, but the new preview is now starting to have the same issue with more frequent “I can’t do that Dave”-style answers, though it’s still mostly circumventable with enough prompt massaging, but it is getting harder.

In a year, I don’t see anything but self-hosted models usable for anything not corporate glitz if trajectories hold, so fuck all that AI safety.

cwagner,

Nothing that runs on my GPU / CPU comes even close to GPT 3.5, GPT4 is not even in the same universe, and that’s with them running far more slowly.

cwagner,

That’s an interface for models. Which model did you use?

cwagner,

Which model are you talking about?

cwagner,

Oh, faster is easy. GPT 3.5 is also far faster than GPT 4. Faster at quality replies is the issue.

cwagner,

I do not believe any 7B model comes even close to 3.5 in quality. I used LLama V1 64B, and it was horrible in comparison. Are you really telling me that this tiny model gives better general answers? Or am I just misunderstanding what you are saying?

cwagner,

Prompt:

I’m currently trying to show on the Website Beehaw, that certain LLMs are far superior in writing than others. Examples of what bigger models do better than smaller ones: *

Mistral-7B-Instruct-v0.1

  • ntire articles* vs Headlines Descriptions vs Product titles *Bul

GPT 3.5-Turbo doesn’t support completion as it’s for chat, so I used an even worse one, text-davinci-003 which is far behind state of the art.

Bigger models are able to handle more complex and detailed tasks with ease

  • Bigger models are better suited for natural language understanding and text processing
  • Bigger models are able to learn more accurate representations of context, thus improving the precision of the output
  • Bigger models can process data more quickly and efficiently, saving time and processing power when large volumes of data are used
  • Bigger models can better recognize more subtle nuances in language, which allows them to produce more accurate results
  • Bigger models are able to use more sophisticated algorithms, resulting in a more comprehensive and deeper understanding of the data being used

Mistral 7B might be okay for some very specific cases, but it’s not comparable to proper models at all.

edit: gave it a second chance, it’s a bit better (at least no complete nonsense anymore), but still terrible writing and doesn’t make much sense

Paraphrasing The ability of a language model to generate text that has a similar meaning to the original text is called paraphrasing. This is a very common problem in natural language processing, and many LLMs are designed to be able to paraphrase text. However, there are some LLMs that are particularly good at paraphrasing, and these models are often preferred over smaller models because of their ability to generate more varied and unique text. Examples of LLMs that are known for their paraphrasing abilities include GPT-2 and transformers. These models

cwagner,

Yes.

cwagner,

Crazy, the news almost took hackernews down when it broke. MS also was taken by surprise, and today 3 lead researchers resigned. Currently only speculation and no one really knows what’s going on.

cwagner,

Maybe don’t sling mud like you are Fox News, even if you are a fan.

cwagner,

I still wonder what’s so different with my circle that so many of them use signal. I mean sure, none use only signal and all also have WA, but almost everyone I know, which is people from Flensburg (border to Denmark) to RLP (center-south), aged late 20 to 70+, use signal. Less than 10% don’t.

cwagner,

That some think feminism means female supremacy

Doesn’t help when some feminists think that as well. We have a famous and influential feminist here in Germany (Alice Schwarzer), who did a lot for feminism in 70s and 80s, but now is a reactionary who stands for TERF, censorship, mass surveillance, female supremacy, and “anyone who doesn’t like me is anti-feminist”.

cwagner,

I usually buy games for under 10, or kickstart them for more. This one’s pitch sounded awesome enough that I backed it on fig originally, and now bought the season pass ;)

cwagner,

Here in Germany in my circle (which has people from mid-twenties to 60+, from the North to the center), most people use Signal, with Telegram being a rare outlier. WhatsApp is what everyone uses, though.

cwagner,

If someone thought he was innocent, SBF would probably “well, actually…” them.

cwagner,

Another suggestions: I like Semafor.com for their newsletters (that I subscribe to as RSS via Kill the Newsletter, also selfhostable) because they give this high level executive overview instead, usually with a tiny bit of humor inserted in their image-tag line. It’s mostly not too US centric (and obviously their Africa newsletter is not at all). You don’t usually get bogged down in stuff, while still being able to keep up with what is happening.

cwagner,

Using anti-competitive tatic to try to eliminate a competitor is literally malicious.

To be fair, it sounds less like eliminating, and more like “carving out a niche”, with the competitor being the dominant one. Doesn’t make it great, but at least a bit better ;)

cwagner,

I don’t play AAA games (except BG3 if that has 3 A’s), but my wife plays The Sims, and between some cheap DLCs and mods, it seems like it’s almost free for tons of content?

cwagner,

Listerine used to have alcohol in their mouth wash, so if yours does too, you might want to think about changing to a different one. While there is no sufficient evidence, there’s a chance that it raises your chance of oral cancer. And IMO you probably don’t need the alcohol in it.

edit: 2009 summary, 2019 meta study

cwagner,

I totally do not get why people have to try to prove to themselves that masks do not work.

I see it as a failure of politics. It’s been a while now, but IIRC in the US as well people were told that they shouldn’t wear N95, just like we in Europe were told not to wear FFP2. That they are hard to use and best left to professionals. That having a beard makes them useless. Now there was a decent idea behind it, stocks were low and uncertain, and healthcare professionals needed respirators first. But then production ramped up, and they were everywhere, and suddenly those masks are great and everyone should wear them.

That kind of political messaging has to produce conflict.

cwagner,

I’m a bit in a not-yet-released slump. “Pathfinder: Wrath of the Righteous” gets a new DLC in late November, so not playing that until then. The same company releases their new game “Warhammer 40k: Rogue Trader” in December, so I’m not playing the Beta until then. “Vagrus - The Riven Realm” gets a new DLC some time this year, so also waiting there. “Colony Ship” gets out of EA and releases November 9, so obviously not playing EA either.

I had enough of BG3 for now (finished it only twice, it’s not as captivating as the Pathfinder games) and will only eventually replay it with one of those “make it like actual D&D5” modpacks. Regarding D&D5, I should some time check if there are updated fan campaigns for Solasta that use the higher DLC level cap :D

I did plan to replay Wasteland 2, but that will take quite some time, and as I mentioned, upcoming releases.

So after telling you what I have not been playing, what I currently do play is alternate between my typical fallback games: Stellaris (a game I never once finished despite 1080h of time played) and Civ V with the amazing and mandatory Vox Populi mod.

cwagner,

Went back to wishlist it. I played the demo, but it was pretty buggy. The “Very positive” is encouraging.

cwagner,

Ohhh, one of my favorite all-time games, only surpassed by the successor.

So first off: Wrath of the Righteous is far more approachable. Better tooltips, better tutorial (including a dynamic one that pops up when relevant), better interface. Just in case you feel like more pathfinder. Note that the powerlevel is far higher in WotR because of mythic powers.

Right after the short tutorial, I went to pick up some berries in a spider-infested cave, which wasn’t too bad, just that the spiders have poison that reduce your stats. I

Fun story: This encounter was heavily nerfed (twice actually, once shortly after release, and then overhauled a few months later). It used to be mandatory and much more brutal :D It literally made some people stop playing :D

Consumable items to remove them exist, but for some reason, drinking a potion can fail. I guess your character just spills everything over the floor.

Hah :D So essentially the potion is a spell in a bottle. Wait, not essentially. That’s what it is. It has a caster level that depends on the creator of the potion. So with that caster level when drinking it, the spell casts Remove Curse against the DC of the curse. And that is what can fail.

After those two experiences, like an hour into the prologue or act 1, I was ready to get fucked at every turn

Sounds like you didn’t encounter the overleveled undead random encounter on the western side ;) Owlcat generally overhauled quite a lot of the encounters to remove difficulty spikes.

My only other gripe with the game is the Kingdom Management.

That is a very common complaint. I liked it, and the time pressure it added. Though only for the first few hundreds of hours (1283h played in total, and without beta testing), later I did what you did.

cwagner,

decided to start with Kingmaker

Yeah, it was a good idea, as usual, for the reasons you mentioned. I just wanted to shill the successor a bit just so you won’t skip it because some things with PF:KM irked you ;)

(I’m playing on the recommended difficulty for someone new to the Pathfinder system)

You are smarter than 90% of people playing the game ;) It’s crazy, their games have super detailed difficulty settings, and presets with explanations. And yet people go, play on the brutal because they like Dark Souls and then complain it’s too hard. It’s like there’s some mental health pandemic with gamers. /rant

but you can pretty comfortably win against

Besides the difficulty, TB vs RTwP also influence this. I play on Core (actual setting in WotR, requires manual changes for KM, essentially as close to tabletop as possible; then some mods to bring it even closer to TT), but a lot of fights I could not do with RTwP, but they become far easier once I play the game in the way the system was designed.

cwagner,

It was painful ;) TB actually started as a mod, that was later with permission taken by Owlcat and improved for the enhanced edition. During the development, there was almost an even split between devs preferring TB and RTwP, and RTwP barely won. So with the mod having laid some groundwork and it being extremely popular, it was easier to pitch the addition.

And yeah, I’m an old school RPG fan, I love TB. I actually stopped playing RPGs just before RTwP became popular with the Infinity Engine games, and only returned to them with NWN2.

  • Wszystkie
  • Subskrybowane
  • Moderowane
  • Ulubione
  • •
  • random
  • polska
  • memy
  • astronomia
  • hosting
  • krotkofalarstwo
  • ukraina
  • Geopolityka
  • kopnij
  • Wszystkie magazyny
    •