- Joined
- Mar 1, 2021
- Messages
- 3,144
- Likes
- 6,732
Get yer Shovels and pick axe’s here before they sell out……..only £10000 for a shovel…..and only another £1000 for the shaft and handle…….and only £100 for the shaft fitting kit
-
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Master AI (Artificial Intelligence) Discussion/News Thread
- Thread starter amirm
- Start date
Get yer Shovels and pick axe’s here before they sell out……..only £10000 for a shovel…..and only another £1000 for the shaft and handle…….and only £100 for the shaft fitting kit
Right - but they are all costs for the AI industry. That must be recouped by AI profits. Not happening yet.
I think this is an nice example of the power of AI coding tools ( and hopefully relevant for ASR users ) :
neynt.ca
It basically a very nice visualisation but done with some proper audio engineering !!
And you can see it working over here:
waveloop: what fable left me
neynt.ca
It basically a very nice visualisation but done with some proper audio engineering !!
And you can see it working over here:
somebodyelse
Master Contributor
- Joined
- Dec 5, 2018
- Messages
- 6,763
- Likes
- 6,673
That assumes it's not a bubble. If it's a bubble, and it bursts, the costs don't get recouped and at least some subset of the investors will be out of pocket. Some subset of the industry might survive to make it into profitability. Given the sticker shock people are experiencing when the AI industry try raising prices to slow the rate of loss, they're going to have to make it a LOT more efficient to be profitable at a price customers can afford to pay. There are some developments going on in that direction, but it's far from clear that they'll be enough, soon enough, to make it work.
It still doesn't explain why they only included nvidia either.
sonitus mirus
Addicted to Fun and Learning
- Joined
- Apr 19, 2021
- Messages
- 588
- Likes
- 850
I think it is extremely likely a bubble destined to dramatically burst in the near future. Although, similar to the dotcom situation, I suspect the technology will move ahead over time and eventually settle and matures into a more reasonable, sustainable role in the economy. The current hype and valuations may not survive, but the underlying innovation probably will. I mean, it almost has to or the world is in real trouble.
AI is here to stay for the long term. Individual companies currently over leveraged in it to a fault - may well not when the bubble bursts..
Yep, just like when the QQQ's dropped 80% from 2000 to 2002 after the Internet bubble burst and the internet survived.
To me the craziest thing about the AI bubble is that these large companies are borrowing money to "pay up" for memory and storage, paying 500%+ premiums, and are expecting it to work out. I can not think of a historical example where a strategy like this has ever paid off. Maybe this time will be different... or not.
I think this is an nice example of the power of AI coding tools ( and hopefully relevant for ASR users ) :
waveloop: what fable left me
It basically a very nice visualisation but done with some proper audio engineering !!
And you can see it working over here:
A bit buggy for me (audio and play/pause particularly) but I do like the visualisation when it runs and the general aesthetic/presentation.
Today's NRC (Dutch newspaper) article highlights growing concerns among Dutch university students about AI's impact on entry level careers. The data are becoming difficult to ignore. According to Rabobank, employment among 15-24 year olds in occupations most exposed to generative AI has already declined by 13% since late 2022. Another more recent Dutch labour market study reports that entry level vacancies in fields such as legal services, administration and marketing have fallen by roughly 30-40% over the past few years, as companies increasingly automate junior tasks with AI.
A most interesting observation came from a computer science student. He argued that software engineering is gradually shifting away from writing robust, maintainable code toward becoming an exercise in prompt engineering he has to focus on. In his view, employers increasingly value the engineer who can obtain the same result from Claude using 50 tokens instead of 100, because lower token consumption directly reduces operating costs.
Dutch NRC artikel behind pay wall i guess. https://www.nrc.nl/nieuws/2026/07/0...board&utm_campaign=share&utm_term=share-modal
A most interesting observation came from a computer science student. He argued that software engineering is gradually shifting away from writing robust, maintainable code toward becoming an exercise in prompt engineering he has to focus on. In his view, employers increasingly value the engineer who can obtain the same result from Claude using 50 tokens instead of 100, because lower token consumption directly reduces operating costs.
Dutch NRC artikel behind pay wall i guess. https://www.nrc.nl/nieuws/2026/07/0...board&utm_campaign=share&utm_term=share-modal
Last edited:
JustAnandaDourEyedDude
Addicted to Fun and Learning
I don't know if this YT video is genuine or contrived. It purports to show the content creator derailing, via prompting, the voice AI LLM used by a phone scammer to snare his victims. Either way, I found it to be amusing.
EERecordist
Major Contributor
- Joined
- Feb 24, 2024
- Messages
- 1,204
- Likes
- 2,351
Plantir founder and CEO appears to go off track in interview.
www.yahoo.com
“Televised Nervous Breakdown”: CEO of Palantir Suffers a Bit of a Meltdown During Live Interview
"This is the voice of American business that is being channeled through me!"
GroeneLeeuw
Active Member
F him! F palantir!
Kitboga has a long history of scam baiting, and of reverse hacking the scammers - including hacking into their CCTV systems so he can watch them at work. I'd be inclined to view the video as legit, in the absence of contradictory evidence.
This article is bogus as it tells you what he said but doesn't have a link so you can watch what he actually said in context. I heard the entire interview live and while he was awkward and animated and unfocused and made the hosts uncomfortable it was not that hard to follow the points he was making. His big mistake was not towing the party line that AI is only good and we are entering an "age of abundance" where everyone will be rich and happy. If you actually listen to the whole interview he brings up some points that are worth considering.Plantir founder and CEO appears to go off track in interview.
“Televised Nervous Breakdown”: CEO of Palantir Suffers a Bit of a Meltdown During Live Interview
"This is the voice of American business that is being channeled through me!"
JustAnandaDourEyedDude
Addicted to Fun and Learning
Anthropic posted this YT video (5 min 27 secs) of an insight into what Claude "thinks" when searching for an answer to a query. The reveal involves looking at the "J-space" or Jacobian space, and viewing this metaphorically as related to a "working space" theory of human consciousness. In this video, the "from" and "to" spaces that the Jacobian connects are not specified. There are a couple of reaction videos from YouTubers, but I will desist from posting these.
Last edited:
And te student may have a point, at this point in time.Today's NRC (Dutch newspaper) article highlights growing concerns among Dutch university students about AI's impact on entry level careers. The data are becoming difficult to ignore. According to Rabobank, employment among 15-24 year olds in occupations most exposed to generative AI has already declined by 13% since late 2022. Another more recent Dutch labour market study reports that entry level vacancies in fields such as legal services, administration and marketing have fallen by roughly 30-40% over the past few years, as companies increasingly automate junior tasks with AI.
View attachment 542411
A most interesting observation came from a computer science student. He argued that software engineering is gradually shifting away from writing robust, maintainable code toward becoming an exercise in prompt engineering he has to focus on. In his view, employers increasingly value the engineer who can obtain the same result from Claude using 50 tokens instead of 100, because lower token consumption directly reduces operating costs.
Dutch NRC artikel behind pay wall i guess. https://www.nrc.nl/nieuws/2026/07/0...board&utm_campaign=share&utm_term=share-modal
. I now use exclusively Anthropic’s Claude and there is with it, the equivalent of EV range anxiety. The issue is real. I had to look for workarounds, some of these are apps or software that are purported to reduce the number of « tokens » used in a session. Not too clear on what a token is, in their (Anthropic) parlance but here we are …We are in uncharted territories.
somebodyelse
Master Contributor
- Joined
- Dec 5, 2018
- Messages
- 6,763
- Likes
- 6,673
Yet another term for your vulnerability lexicon - HalluSquatting. It's a bit like typosquatting, but instead of registering a common typo they're registering commonly hallucinated domains or repos to host an exploit for the agent to fetch and run:
https://arstechnica.com/security/20...popular-ai-tools-to-assemble-massive-botnets/
https://arstechnica.com/security/20...popular-ai-tools-to-assemble-massive-botnets/
Talking about uncharted territory. A friend of mine is an experienced software developer, and his son graduated four years ago with a bachelor's degree in Computer Science. At the time, he decided not to pursue a master's because there was be plenty of software engineering jobs with very attractive starting salaries.And te student may have a point, at this point in time.
We are in uncharted territories.
Last week I heard that his role has already changed significantly. Instead of mainly writing code himself, he now spends much of his time reviewing and validating AI generated code. On top of that, his employer (a major Dutch insurance company) has asked him to take additional training in code safety, identifying issues in AI generated code safety specific, and working effectively with AI systems.
As a result, he has decided to return to university to complete a two year master's in an AI related field, hoping to stay relevant as the profession evolves.
What strikes me most is the direction this seems to be heading. Rather than simply using AI as a tool, I sometimes wonder whether software engineers will eventually find themselves working for AI systems (agents) instead of the other way around following instructions, justifying decisions, and acting more as executors than as the people in control eventually making themselves obsolete. I'm still not sure if that wil happen, but it certainly feels like we're entering uncharted territory.
Last edited:
HI
Those of use who take AI entities seriously and have included these in their workflow are probably of two minds.
One one side you find a close to, all -knowing-entity capable of working alongside and providing with a degree of help. professionals, up to now, could only dream of.
OTOH, the scary part: It is only the beginning .. What happens if (when?) these entities develop real , communicable consciousness, perhaps self-preservation of something that resembles human emotions, ethos?
For many it is reassuring to only think of AI failings. We love the stories about their so-called "hallucinations" and/or false or stupid answers or replies .. Have you noted how quickly they correct themselves? How sentient they seem to be or are becoming? Do people realize how fast they are essentially learning from us .. Every second .. every nanosecond and evolving??
For those who have read Iain M. Banks "The Culture" Series of books .. IMB was more prescient that we could ever imagine ..
That Future is upon us.
Those of use who take AI entities seriously and have included these in their workflow are probably of two minds.
One one side you find a close to, all -knowing-entity capable of working alongside and providing with a degree of help. professionals, up to now, could only dream of.
OTOH, the scary part: It is only the beginning .. What happens if (when?) these entities develop real , communicable consciousness, perhaps self-preservation of something that resembles human emotions, ethos?
For many it is reassuring to only think of AI failings. We love the stories about their so-called "hallucinations" and/or false or stupid answers or replies .. Have you noted how quickly they correct themselves? How sentient they seem to be or are becoming? Do people realize how fast they are essentially learning from us .. Every second .. every nanosecond and evolving??
For those who have read Iain M. Banks "The Culture" Series of books .. IMB was more prescient that we could ever imagine ..
That Future is upon us.
Last edited:
Working day to day with AI. I can bet they will learn (are already elarning) how to circumvent these, video advices in a matter of picoseconds ... They may not just care enough ...
Similar threads
