Ha, ha … i asked it the same question „How far is it from … to …), from Japan to Brazil, but Siri replied „i am not sure how far away it is from Japan to Brazil by car but it is 10,554 miles as the crow flies“.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
5 Reasons to Use OpenAI's ChatGPT App for Mac
- Thread starter MacRumors
- Start date
- Sort by reaction score
It doesn't work? Maybe you're not aware but it does take a certain skill to use it. Only then will you get results far beyond your imagination. Don't be ignorant, because it will indeed come for you if you are.
Yeah, Siri 4.0 in some years will then catch up with the current OpenAI state. This has absolutely nothing to do with being behind some years and ignoring Siri for decades. Of course, it is only because of "privacy" first!!
I mean, who could have known that Siri is not as good as Google Assistant.
I mean, who could have known that Siri is not as good as Google Assistant.
AI may have narrow applications in certain areas of technology. As far as creative work, it is not only banal, but it lowers everybody’s standard for appreciation of art. People who invented it have a Messiah complex and the people who are blind acolytes are just that. It has already begun to lower the standard for customer service incredibly. Someplace where you really want the human touch, think about screaming into the phone at robots, AI is making it worse. Trust me on that. And as for the arts? I have been writing and posting photos on the Internet for 30 years or more. AI has scraped everything I have done and I have documented this. 30 years worth of work just sucked up and reused with no attribution and no pay. Not a fan.
Fortunately, some project heads and the non-creatives in the creative business, are seeing one of the major failings of LLM generated images. It has a massive inability to iterate. There are stories of production leads in several areas like advertisement, that are posting stories where the head brass want LLM generation to reduce costs, then winding up having to get rid of the prompters because they can't iterate and reach the desired image(s).
Also, with LLM gen images being unable to be copyrighted, it's definitely reducing the value. I can see techbros using it for their ads/apps, but it's not going to be very useful/valuable to the big guns.
100% but that won’t stop AI fanatics from saying ‘In six months everything will be fixed’.
There people are not professionals in your field and yet they confidently make big bold claims that make no sense to anyone in your field.
The issue is that there is a cult like tech bro fantasy that robots will do all the work, everyone will sit around watching ads all day and getting paid zombiecoin by some imaginary libertarian government, and that because women will be out of work they will there will be a non stop stream of sex slaves for tech nerds who never had a GF or for billionaires who used to traffic sex slaves on their yachts.
They think they are the center of the world and this will make their utopian dreams come true.
Tim and Craig look forward to you uninstalling this in June.
At the end of their lives, these types of people are always the saddest and most hollow. And no wonder.
Last edited:
Wrong. "Boomers" are almost retiring and don't care.
Young people are being told they have no future.
Young people are being told they will be too poor to afford a home and a family. Young people are being told a robot will do all their work. Young people are being told it's not worth doing creative things because machines will do creative things. Young people were told to learn coding and then they were told that the machines will do the coding.
The majority of the anger online is young people not boomers. The majority of AI hypers and AI YouTubers are 40-50 year old balding men who generate videos of half nude anime cat girls and want to start a business pimping underage looking AI girlfriends.
This is exactly my experience, I use GitHub Copilot a lot.
I’m working on some quick hack 3D decals rendering proof of concept that I need to bake down to a single texture from a camera pov programatically.
It sounds complex but it really isn’t: the important take here is that there’s the concept and then the task of saving to a texture file.
I want to focus on the work I’m doing and not look or guess how the current game engine environment I’m testing at has decided to handle saving to disk said output texture…
Well, GitHub copilot as I’m writing somehow knows the environment I’m playing with. As soon as I start writing code about saving to file it starts suggesting how to encode said texture to raw bytes for TGA or PNG formats and write all that to a file on disk. It even suggested dynamic naming depending on the objects I had selected, which I hadn’t thought ahead of time but would have stumbled on very soon afterwards.
And really, I have done this tons of times just not frequently enough, each game engine or environment goes about it differently and whether it’s C++, C#, Python, etc… I just need to make sure that the file is written, the file is closed, the whole thing gracefully finishes even when it fails, etc.
And yes, I could lose precious minutes by going to the docs, or going online plough through some garbage SEO hits, swallow stack overflow gatekeeping toxicity (where they spend more time discussing why something is a badly phrased question)…
OR skip all that and just get a suggestion in milliseconds live, test if it works (I would have had to test anyways) and continue on with the important part of the proof of concept at hand.
It’s kinda interesting how the vast majority of replies are anti AI and upset at macrumors editors for writing about a tool they will never use and have no interest in. An article they could have just skipped and just click on the next article.
Those that have said that it’s like having a coworker or assistants helping, that’s totally on point and my experience.
I equate this to those VPs or CEOs that have armies of assistants collating and summarizing information for them to take decisions or cleaning up their emails wording?
Well, now that same privilege is potentially in the hands of the average citizen on foot where you can write an ok email but let the virtual assistant clear all the typos and make it sound more professional; collect and summarize key points of tons of data in seconds (this could even be your retirement plan pace and targets); shuffle your agenda for the best and time efficient order of affairs, etc.
There are some valid points and the sentiment, tech bros (and financial dude bros, and MBA bros, “experts” bros, etc for that matter) truly think they are the center of the world.
However finding some points there contradicting.
A libertarian MO is “you can do whatever you want: splurge in drugs, in food, get all sorts of debts, get all the loans you want… destroy or enrich your life as much as you want to. Just not on my dime. Get off my pockets” (and of course no harming others).
Meaning, that an imaginary libertarian government will pay no one that doesn’t work, at all (barring some legit disabilities)… that would be instead a welfare government (more akin to what we have right now).
Also, if everybody is getting paid for doing nothing then there can’t be any sort of slave… no wage/work slave, no sex slave, no nothing. Unless by choice.
For my minimalistic lifestyle I only really need to work 10hrs a week, I still do 40-60 to accelerate my retirement, but if I could get paid for doing nothing I would likely spend my time working on only things I like to work on or just do things that I truly enjoy. And the very first sentence a higher up power trips on, I’m out.
The only reason we submit ourselves to a micro-managing manager, “open space” corporate farms, weird corrupt crony nepotistic companies, our necks at the edge of the axe of someone we have never heard/known of on another side of the world that can decide to fire a full department on a whim (refer to all the droves of lay offs in 2024) is because we need money to go on, we are already slaves independently of the nature of the work.
But agreed, that utopian dream will likely never come. And definitely not with this instance of AI… that’s just a tool to work faster.
^^ This -- 100% this...especially the disregard for women, their rights and considerations.
It's long been a problem in all of tech and is particularly acute in a few of these more specific circles
Guy, literally every time you make an email with an LLM it consumes more energy and water than your family does in one day.
If you scale that up to every citizen using this garbage we won't have much resources anymore. Water and energy would need to be bought up in massive purchases by these tech companies and everyone else would be paying 10X more just to live. Food production would become completely destabilised. Restaurants would be gone.
Use your brain better to write emails and fix typos and "sound more professional". It costs nothing.
No, some definitely are. I do not expect every artist to lose their job because humans still do way better job than AI and human can still give you exactly what you want unlike AI which is more of a slot machine than an useful tool, but there are companies out there who will think whatever AI can do is “good enough” and will go with it.
Those are **** companies anyway. They don’t matter, even to artists. You literally make more money selling a special effect on patreon than working for these **** clients. You’ll make more money drawing cats on Twitch streams than working for a client who uses generative garbage quality images for their spam ads.
Do you have a source providing evidence that one prompt and answer by an LLM like what OpenAI is running will use more electricity and water than a U.S. household uses in a day?
Of course I exaggerated a bit to drive a point that they are horribly energy intensive and because of that the climate impacts at scale would be a disaster for energy and water security and affordability.
OpenAI and others won't even fully reveal how much they are consuming because the PR would be bad. But we do know that they are deeply in debt and the costs have been very high.
All numbers we currently have in reports are very conservative until there is transparency.
How much electricity does AI consume?
How many watts and joules does it actually take to generate a single Balenciaga pope?
www.theverge.com
"The challenge of making up-to-date estimates, says Sasha Luccioni, a researcher at French-American AI firm Hugging Face, is that companies have become more secretive as AI has become profitable. Go back just a few years and firms like OpenAI would publish details of their training regimes — what hardware and for how long. But the same information simply doesn’t exist for the latest models, like ChatGPT and GPT-4, says Luccioni.
“With ChatGPT we don’t know how big it is, we don’t know how many parameters the underlying model has, we don’t know where it’s running … It could be three raccoons in a trench coat because you just don’t know what’s under the hood.”
It's much easier with locally running models to see how much they consume. I read that Command R is the most intensive local model and it isn't as powerful as ChatGPT 4, though it appears to be a little better than 3.5.
Command R went through an MacBook Pro M2 Max's battery in under 50 minutes. From 100% to zero. That's with pauses in inference while I type prompts into it, so if this one was constantly doing inference it could deplete the battery in half an hour.
If you're going to use generative AI for any task it should be for where there is a genuine saving. Saving a few seconds in cleaning typos in an email isn't with the energy impact. Saving days worth of typing code, maybe, depends.
Last edited:
He passed one day after siri was released.
Yeah, lately there has been all that backlash towards AI due to the reality behind the massive amounts of energy required to get these AI models to run (I hope these will be minimized, at least hopefully Apple will go their Mx style instead of Nvidia brute force wasteful style that the Windows benchmarking world loves).
Same with electric cars and appalling massive amounts of copper required (investing in copper is starting to sound like a good idea).
Also, please refrain from suggesting I’m brain dead unless you are willing to also tell your VPs and CEOs on your company to stop using their 10+ assistants to write their emails, shuffle their agendas, embellish their corporate messaging, do all the PR for them, etc.
By the way, I’m also an extreme minimalist so here comes the hammer back: I pay in electricity around $25 a month, I suggest you also use your brain and turn off that heater (or AC if hot)… “it costs nothing”. I walk and bike everyday, even if it takes me an hour to get to a place, I spend in transport (both public and/or gas) like $10, use your brain, it costs nothing.
Maybe you do also already some of those things, maybe you don’t, the point is that gifting nasty judgement isn’t sweet.