Welcome to Incremental Social! Learn more about this project here!
Check out lemmyverse to find more communities to join from here!

kbin_space_program

@kbin_space_program@kbin.run

This profile is from a federated server and may be incomplete. Browse more on the original instance.

kbin_space_program ,
kbin_space_program ,

The article I posted references a study where chatgpt was wrong 52% of the time and verbose 77% of the time.

And that it was believed to be true more than it actually was. And the study was explicitly on programming questions.

kbin_space_program ,

I haven't had need to do it.

I can ask people I work with who do know, or I can find the same thing ChatGPT provides in either la huage or project documentation, usually presented in a better format.

kbin_space_program ,

I never said that.

I said I found the older methods to be better.

Any time I've used it, it either produced things verbatim from existing documentation examples which already didn't do what I needed, or it was completely wrong.

kbin_space_program ,

If something is going to blow up, its much better to happen on a test stand than on an actual product or test launch.

Best case would be doing the math beforehand, as they Didn't do with the flame trench iterations until the water pump system was added. And we know that because other people on youtube did do the math and determined even the special high temperature concrete from NASA wasnt going to be enough by itself.

kbin_space_program ,

Dude, the entire pad was gone. People in the "safe" zone had concrete raining down on them and the rocket itself was severely damaged from the takeoff.

If they had done the math before that, they would have never attempted that launch.

kbin_space_program ,

Google search isnt a hallucination now though.

It instead proves that LLMs just reproduce from the model they are supplied with. For example, the "glue on pizza" comment is from a reddit user called FuckSmith roughly 11 years ago.

kbin_space_program , (edited )

You're right that they arent hallucinations.
The current issue isn't just summarized web page, its that the model for gemini is all of reddit. And because it only fakes understanding context, it takes shitposts as truth.

The trap is that reddit is maybe 5% useful content and the rest is shitposts and porn.

kbin_space_program ,

Here is one news source reporting on the fun:

https://www.dailydot.com/debug/google-search-results-reddit-pizza-glue-cheese/

It was an actual shitpost. I had originally assumed the same as you, given that there are a few bloggers and youtubers who go through the tricks used for food photography.

kbin_space_program ,

It is, but it isnt applicable in at least the glue-pizza situation as the probable source comment has been found on reddit.

A better use of the term might be how when you try to get Bing's image creator to make "Battletech" art, you just mostly get really obvious Warhammer 40k Space Marines and occasionally Iron Maiden album art.

kbin_space_program ,

There are a lot of people, including google itself, claiming that this behaviour is an isolated and basically blamed users for trolling them.

https://www.bbc.com/news/articles/cd11gzejgz4o

I was working on the concept of "hallucinations" being things returned that are unrelated to the input query, not directly part of the model as with the glue-pizza.

kbin_space_program ,

A Google spokesperson told the BBC they were "isolated examples".

Some of the answers appeared to be based on Reddit comments or articles written by satirical site, The Onion.

But Google insisted the feature was generally working well.

"The examples we've seen are generally very uncommon queries, and aren’t representative of most people’s experiences," it said in a statement.

It said it had taken action where "policy violations" were identified and was using them to refine its systems.

That's precisely what they are saying.

kbin_space_program ,

Or Musk decides that you don't need some part of your brain. Or worse, rents it out as server space.

kbin_space_program ,

I believe the SAG(Screen Actors Guild) just had a big strike about AI likenesses without proper compensation.

kbin_space_program ,

Open AI seems to have dealt with their massive legal issues so far by signing caught red handed agreements.

kbin_space_program ,

It has happened before. They just swept it under the rug and blamed the client.

A user was setting up a new laptop and synced an empty folder with google drive, intending to download accounts data to their machine. It bugged and treated the empty folder as the master and began erasing the drive contents.

After two weeks of pestering google, they relented and pulled from their backups they swore they didn't have and didn't exist.

kbin_space_program ,

The thing is that the AI text is atrocious and vapid. It takes up a lot of space and says

"Laser printers are better in every way minus full color than inkjets, but are bigger and more expensive than inkjet printers."

The trick is that AI took 12 paragraphs and using a list incorrectly to do it instead of a sentence. And the editor calls it out for that.

kbin_space_program ,

The entire fundamental basis of the CCPs economic miracle is that intellectual property is not respected.

kbin_space_program ,

One theory Ive read about is that they probably merely iterated on the (likely middle eastern, probably egyptian or persian) invention of "greek fire". Since its only a single ingredient difference between one of the known formulations of it.

And the Europeans made gunpowder useful with the invention of pearled gunpowder, which made it possible to predict burn rates and slow them down for cannons, allowing for bigger and more potent cannons. Anecdotally, there is documentation of an Ottoman diplomat pleading to a Chinese one that "the Europeans never learn to make gunpowder"

They also claim to have invented chariots, despite using the Sanskrit word for chariot.

kbin_space_program ,

In short:

Sinovac, "their" covid vaccine, was stolen from Canada when Canada offered to help them produce a vaccine. Signed an official treaty and everything. Canada did the hard work that they literally couldn't, and still can't, do. Then they took the vaccine and claimed it was entirely their own work.

Huawei's 5G development miracle is 100% stolen technology from Nortel. They did no work of their own other than work so shoddy even the UK Tories didn't want to touch it.

These are just the two highest profile ones that go right to the leaders of the CCP.

kbin_space_program ,

They might be using it how Tesla markets it.

kbin_space_program ,

https://www.youtube.com/watch?v=tlThdr3O5Qo

A 2 minute long video of a tesla showing self driving in traffic where the driver isnt touching the wheel. From Tesla and on their official channel.

Edit: from at least 5 years ago.

kbin_space_program ,
kbin_space_program ,

Tesla's official channel from 5 years ago says otherwise. Its nice of them to put up incriminating material that a layman can find in 5 seconds.

kbin_space_program , (edited )

The state of California does not agree with you.

https://www.reuters.com/business/autos-transportation/california-regulator-claims-tesla-falsely-advertised-autopilot-full-self-driving-2022-08-05/

Edit: Tesla's own website still advertises Full Self Driving as an enhancement of autopilot.

kbin_space_program ,
kbin_space_program , (edited )

There are massive fundamental driving differences between the provinces.

For example, you can instantly tell an Albertan driver. They don't know how to do hills, corners, or worst of all, hills with corners.

kbin_space_program ,

If youre talking about car insurance, comparing rates by themselves is irrelevant. Alberta has relatively cheap insurance, but BCs basic coverage is more than ten times better for only about 30% more cost.

kbin_space_program ,

That price is enough, provided the company doesn't get infected by MBAs at the management level.

kbin_space_program ,

Yes.
If SO1 dies a decade or two before SO2, SO2 isn't expected to be faithful, why should SO1 in the unlikely scenario that an afterlife exists?

kbin_space_program ,

The Vasa comes to mind.
https://en.m.wikipedia.org/wiki/Vasa_(ship)

USSRs space shuttle, although technically that didnt fail on its own.
https://en.m.wikipedia.org/wiki/Buran_(spacecraft)

The failed Chinese attempts to use stolen SpaceX plans that resulted in at least one town that we know of being destroyed with toxic smoke and fire.

Virgin's space program.

Boeing's space program. Increasingly, the 737-MAX too.

Amazon's space program.

Musk and the stupid decisions made around Starship, that resulted in over a year of setbacks since the launchpad had a woefully inadequate flame trench that everyone called out long before the first two failed launches.

The Titanic.

Related to the last one: OceanGate and the Titan.

When the Athenians starts the Peloponnesian War by invading a different city state than Sparta(to ostensibly use that city's military against Sparta), but screwed it up, creating the conditions for them to lose said war.

Instagram Advertises Nonconsensual AI Nude Apps (www.404media.co)

Instagram is profiting from several ads that invite people to create nonconsensual nude images with AI image generation apps, once again showing that some of the most harmful applications of AI tools are not hidden on the dark corners of the internet, but are actively promoted to users by social media companies unable or...

kbin_space_program ,

Nice pipe dream, but the current fundamental model of AI is not and cannot be made deterministic. Until that fundamental chamge is developed, it isnt possible.

kbin_space_program ,

https://www.sitation.com/non-determinism-in-ai-llm-output/

AI can be made deterministic, yes, absolutely.

The current design everyone is using(LLMs) cannot be made deterministic.

kbin_space_program ,

I love your wishful thinking. Too bad academia doesnt agree with you.

Edit: also, I have to come back to laugh at you for trying to argue that the almost random nature of software random number generators is deterministic AI.

kbin_space_program ,

You can make it more deterministic by reducing the acceptable range of answers, absolutely. But then you also limit your output, so thats never really a good use case.

Randomness is a core functionality of not just LLMs, but the entire stack that has resulted in LLMs. Yes you can get a decently consistent answer, but not a deterministic one. Put another way, with LLMs being at max constraint, you can ask them to add 1+1. You'll usually get 2. But not nearly always.

kbin_space_program ,

No, thats not how it works. Ive already explained and posted links as to why.

Why is Matrix mentioned more often than XMPP in self hosted forums?

I'm looking into hosting one of these for the first time. From my limited research, XMPP seems to win in every way, which makes me think I must be missing something. Matrix is almost always mentioned as the de-facto standard, but I rarely saw arguments why it is better than XMPP?...

kbin_space_program ,

If nothing else, there is space for a competitor to MS teams in the corporate space.

Everyone else is ending up on teams, but no one actually likes it.

kbin_space_program ,

No, not always. I know of a very major firm that uses google suite for everything but chat and video calls. They use MS Teams because its just that much better than google's alternative. From the chats Ive had, the issue with Slack there is that someone high up in their IT stack hates it.

'Vortex Cannon vs Drone' - Mark Rober shows off tech from a "defense technology company that specializes in advanced autonomous systems". That seems bad

I've enjoyed Mark Rober's videos for a while now. They are fun, touch on accessible topics, and have decent production value. But this recent video isn't sitting right with me...

kbin_space_program ,

You're approaching the issue incorrectly, because you're omitting cost.

For example: Russia is using suicide drones that cost a few hundred to a few thousand dollars each.

It's not economically(or logistically) viable to fire a few hundred rounds of ammunition at every drone.

Firing a several thousand dollars worth of bullets at a missile works because the missile is at least several hundred thousand.

kbin_space_program ,

Eh, the comparison is valid when the opponent can throw ten to upwards of several thousand drones at you for the cost of one countermeasure.

kbin_space_program , (edited )

Mostly i agree.

However:
Drive by wire is something that they should have implemented decades ago. It is a ridiculously proven and safe technology. Edit: just not in the Cybertruck. Or GM.

kbin_space_program ,

Oh the Cybertruck is an absolute Deathtrap.

First off, the wire control systems in aircraft are absolutely reliable, in part, yes, because of the backups and redundancies. I figured I didn't need to say that.

Because the cybertruck is much worse than you know, based on what you wrote about things going wrong.

You know how sane design has different systems with their own cable bundles, so that if one bundle is damaged, you lose that system, not Everything dead?

Well the Cybertruck puts everything in one singular bundle. If that gets damaged, everything electronic dies. Which can include your backup mirror.

In short: wire control systems, with redundancies, great, long overdue. Cybertruck pos.

kbin_space_program ,

Google docs is only good for collaboration. Otherwise its insane that MS Office is still superior to it.

Hell, Libreoffice is admittedly slightly behind MS for features, and its still superior to Google Docs.

kbin_space_program ,

And Libreoffice does that better

And, you know, is actually free

kbin_space_program ,

Oh sweet green earth. Why the hell would you ever edit documents on an ipad

You need a job with a sane IT policy.

kbin_space_program ,

Because when it's a rust issue, clearly something else is at fault.

Frankly I'm more concerned about Java's wont fix

kbin_space_program ,

R&D is also a fairly big tax break companies can use.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • incremental_games
  • meta
  • All magazines