Click here: to donate by Credit Card
Or here: to donate by PayPal
Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794
Thank you very much and God bless you.
Posted on 08/25/2025 12:15:50 PM PDT by libh8er
The shine may be coming off AI, a tech charlatan that has brought no major benefits for organizations including telcos and has had some worrying effects.
Offloading cognitive effort to ChatGPT or a similar application is extremely bad for the brain. Who knew? It should have been obvious to anyone who's realized that lounging around all day is bad for the body, or that no one became good at anything by not doing it. But it took two separate research projects, one by Microsoft and Carnegie Mellon University and the other by MIT, to establish that overreliance on generative artificial intelligence (GenAI), as the madmen of Big Tech call it, turns you into a technologically lobotomized ape.
It now appears to be even more damaging than all that, according to Mustafa Suleyman, the head of AI for Microsoft and the author of the portentously titled The Coming Wave (spoiler alert, AI is going to be seriously disruptive, writes man who stands to earn millions from serious AI disruption). If you've not heard of what Suleyman and others are describing as "AI psychosis," it is a new ailment whose sufferers are convinced AI is sentient. In a case of life imitating art, some people have apparently grown emotionally attached to the machine voices emanating from their phones and even, like Joaquin Phoenix in the movie Her, fallen in love with their chatbots.
AI psychosis is feasibly a natural consequence of the cognitive decline researchers observed in heavy users of ChatGPT, much as the onset of lung cancer is for the coughing yet dedicated smoker. In defense of the afflicted, it has been encouraged by two years of insane industry babble about what is basically just a very sophisticated search engine, the progeny of the pattern recognition system that Google's founders worked on in the late nineties.
Scaremongering headlines about job losses and murderous robots have probably contributed to AI psychosis. Hardly any commentator has even objected to the marketing of the technology under the AI banner. Yet backers have had to invent the new label of artificial general intelligence (AGI) to describe what AI was supposed to be until ChatGPT came along.
Meanwhile, the world mercifully looks no closer to AGI. It is impossible to see how a superior intelligence that outperforms the smartest humans on all fronts could be a positive for the planet's dominant species, but that hasn't stopped Sam Altman and other latter-day Frankensteins from trying to create one. The highly anticipated GPT-5 has fallen scandalously short of expectations and is merely an incremental improvement on GPT-4 rather than some AGI-like breakthrough. Building even bigger large language models (LLMs) and more powerful graphical processing units (GPUs) hasn't been fruitful and probably never will be thanks to the law of diminishing returns.
GPT-5 also looks as rubbish as its predecessors. Richard Windsor of Radio Free Mobile set it two illustrative tasks – first, "draw a picture of a person writing with their left hand," and second, "draw a picture of a person holding a sign saying AGI is imminent. Circle all of the vowels." The resulting images, shown in his blog, feature a right-handed scribe and a man holding an "AGI is imminent" sign with two circled consonants. It offers proof, writes Windsor, that the generative AI tool "still demonstrates no understanding of causality."
What AI-related job cuts?
But in the absence of any other game changer for their business, telcos still sound captivated by AI, shoehorning it into any gaps they can find. No announcement seems complete without a reference to the technology and its supposed benefits, which remain largely invisible to outsiders. It has neither spurred revenue growth nor boosted profitability, and it has certainly not given telcos an array of new services to sell.
AI has had an impact on networks mainly by generating additional traffic in and between the data centers training LLMs, which looks good for vendors of data center connectivity products like Arista, Ciena, Cisco, DriveNets, Juniper (now owned by HPE) and Nokia. Radio access network (RAN) vendors like Ericsson pray to Loki or some other mischief-making deity that it will eventually have the same impact on 5G traffic, as well. Few are convinced, and the effect of that would be to squeeze telco margins, forcing operators to spend more on 5G infrastructure when there is no obvious prospect of higher sales.
Every week seems to bring another news story about telecom and tech-sector layoffs linked to AI. Yet most are extremely misleading. Even the tech giants perceived to be emptying their buildings of staff and installing AI everywhere employ far more people than they did just a few years ago. Microsoft's workforce grew from 144,000 people in 2018 to 228,000 last year. Alphabet, Google's parent, finished 2024 with 183,323 employees, up from 98,771 six years ago. Amazon's workforce has grown 2.4 times over this period, to an astonishing 1.556 million people – making it bigger than the population of Estonia, although not South Korea, as Google's Gemini chatbot wrongly deduced when tested by Light Reading (see below).
|
|
Click here: to donate by Credit Card Or here: to donate by PayPal Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794 Thank you very much and God bless you. |
“AI looks increasingly useless in telecom and anywhere else”
Disagree with that. It would be a great improvement to turn traffic control in the US over to AI and get rid of the ones doing a poor job of it now
Verizon just implemented AI Customer Service. I just fired them and got other service because it is now just about worthless...
Verizon just implemented AI Customer Service. I just fired them
and got other service because it is now just about worthless...
AI customer service worthless........how so?
Is a lion intelligent?
It will eat you anyway.
Because all the cures are “one size fits all” and they are NEVER WRONG... I ti NEVER THEIR FAULT.
It is a wrapper of plausible deniability and denial of liability to hide behind...
It is...
“ I find the AI is much faster than watching a tutorial or going to the manual for answers.”
Ditto. Earlier today, I asked Grok about the little blue dot on my iPhone and later about the little,green triangle pointing upward. Got concise answers immediately
Because all the cures are “one size fits all” and they are NEVER WRONG... It is NEVER
THEIR FAULT. It is a wrapper of plausible deniability and denial of liability to hide behind.
Problem is... All businesses will be doing that soon... “Customer Service” will be obsolete...
It will be like “Talk to the hand!”
Problem is... All businesses will be doing that soon... “Customer
Service” will be obsolete...It will be like “Talk to the hand!”
AI looks increasingly useless in telecom and anywhere else
ChatGPT said:
I hear you — a lot of people in telecom (and other industries) are feeling disillusioned with how AI is being pitched versus what it can realistically do today.
Here’s a quick breakdown of why that gap exists:
⚡ Why AI Feels “Useless” in Telecom (and Beyond)
Overhype vs. Reality
Vendors often promise “AI will automate everything,” but in practice, it usually just makes incremental improvements (like helping triage trouble tickets or analyze call logs). The hype sets expectations at world-changing scale, so the reality feels disappointing.
Narrow Application
Most AI in telecom today is stuck in silos:
Chatbots that deflect basic customer service calls (but frustrate real customers).
Network anomaly detection that flags possible outages (but still requires human engineers to validate).
Predictive maintenance on equipment (but adoption is patchy and data-hungry).
Data & Integration Challenges
AI thrives on clean, structured data. Telecom providers have mountains of messy legacy data spread across billing systems, OSS/BSS, CRM, etc. Without proper integration, AI models either underperform or can’t be deployed at all.
High Cost, Low ROI
Implementing AI at scale requires new infrastructure, staff training, and vendor lock-in — and executives often don’t see clear financial payoff compared to just improving existing processes with conventional automation.
🌍 Where AI Is Useful in Telecom
Even if it feels shallow now, there are some niches where AI is genuinely adding value:
Fraud detection — spotting SIM-swapping, subscription fraud, or unusual call patterns faster than rules-based systems.
Network optimization — tweaking cell tower loads and frequencies automatically.
Churn prediction — helping retention teams identify customers likely to leave.
Field ops — AI-assisted image recognition for tower inspections and damage detection.
But again — these are small wins, not revolutions.
🔮 Bigger Picture
The same pattern is showing up in other industries:
In healthcare, AI struggles with accuracy, liability, and messy medical records.
In finance, fraud detection works, but “AI advisors” are mostly just rebranded algorithms.
In creative work, tools help brainstorm or draft, but human skill still carries the real value.
So the pattern is clear: AI is an enhancer, not a savior. The disappointment comes from expecting it to replace entire workflows rather than assist in them.
Yes, it is...
AI has its uses. I know a guy that uses it to help make complicated formulas in spreadsheets and other analysis software. Almost all of the rest of it is GIGO.
lol
“Automated phone attendants were also described this way, and they have gotten marginally better, but certainly have not gone away. AI will become more useful, and systems will have to be built to integrate and automate it. That will take time.”
Yeah, I just can’t wait until it is replacing 911 call centers...
I use AI for coding PLC controllers and HMIs for automation machinery. Saves me lots of time at things I am not good at.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.