It's all doom and gloom. Is this AI wave really going to affect people?
17 replies (most recent on top)
@mq I have been reading this website since sometime last year and I'd say this is the best, and most coherent, comment that I have seen here. A jewel in a sea of sh1t. Thank you.
As a student of history, if VZ is hyping it, they will find a way to sc--w it up.
I think even if token costs go up 5x it will likely be worth it for our software engineers. If they go up beyond that, with AI’s current capabilities in tact then we would have to reevaluate. Already it cannot write all of your code not without you in the loop. I’m excited to see what becomes of AI. I do not think we are on the brink of AGI nor is that term very meaningful. Let’s emphasize the general in AGI. Say it becomes generally intelligent but with all of the documented information in its training weights. It still will require human handholding. Look around, how many humans do you know living truly self architected lives? Very few. Most of us get a job, do what we have to for that job and go enjoy life. Almost no one you work with at VZ lives a fully self architected life. They are a piece in a machine doing what they have to. The only people who are self architecting their lives are going direct to customer in the marketplace with their own business system. So unless we are all not generally intelligent, which I think we are generally intelligent. Then automated general intelligence is not going to be more efficient than humans. One we’re already here. Two we make mistakes so will AGI. Three frankly we are cheaper. You have billions of people in the world that live off of less than 2 USD’s a day. All of them would be cheaper to train up than scaling LLM’s. Every single instance of an LLM can only service X amount of requests. They use effectively load balancing software layers to ramp that up, but it’s not infinite. The data center expansion is largely over exaggerated. These frontier models are running into liquidity problems. There is a finite amount of humans in this world that are willing to take the bet that is AGI, and I get the feeling we are running out. Big tech is already starting to hire back some of their better people that they let go. It’s a pendulum. It swung too far in one direction, and I think the momentum is shifting back the other way in the next 24 months.
They are all cash-burning Machines:
OpenAI is burning approximately $17 billion in cash this year against roughly $24 billion in annualized revenue. That’s spending roughly $1.70 for every $1.00 earned at the total operating level.
OpenAI’s burn rate sits at roughly 57% of revenue in 2026 — meaning losses are more than half of what they bring in.
Anthropic
Anthropic’s compute spending reached $6.8 billion in 2025 across model training and inference — exceeding staff and all other expenses combined. Against projected 2025 revenue of ~$4.5B, compute alone exceeded revenue.
GitHub Copilot was losing up to $80 per user per month for heavy users in its early days while charging a flat $10.
Pure-play AI products are broadly loss-making at the total cost level, and currently subsidized by investor capital.
@d4 AI Tokens are highly subsidized at the moment and all Hyperscalers are operating on huge monthly losses in the millions, and they are all burning through cash exponentially higher than any trickle of revenue they receive in return. Look up the cost of compute versus the cost of Tokens, it is a business that is bound to Fail. Prompt Engineer all you want, nobody is saying that it doesn’t help to do certain things more quickly— but without PROFITS to justify all the Leveraged AI spending it may as well be the Next BUBBLE 🫧 🫧 and it is already well on its way to PoP 💥
As OP @as+1krr7hvtn Posted. As a AI Architect in the industry, being sick and tired of VZ antiquated internal policies, You can either live in denial with every possible excuse “ Hype”, “Grammar tool”, “Glorified Spell Check”, “Ponzi Scheme”, and please come up with any and all future reasoning. If you dont know what prompt engineering is, Claude CoWork, Claudebot, Agentic AI, Gen AI…you can either continue to live in your shell (Reminds me of AOL Dial Up CD’s and here we are with Starlink) ….or LEARN !
At this point, I think that the capabilities AI are overhyped, but its functionality is growing rapidly. It will undoubtedly change business processes, but for the foreseeable future, it will remain a tool to assist workers rather than a full-scale replacement for them.
Seems like another Ponzi Scheme, lots of Hype and promises about [ Future ] gains without present use case scenarios that have produced any Profitability. A ton of money being poured down the drain in infrastructure without bringing in any actual Profits, and getting more expensive than having actual employees so let’s see how long that Charade can last. So far it’s just a money grab by the tech titans while putting a lot of people out of work.
Yes, in Finance department that ensures balance of usage and terms of service, AI can replace us. Using AI to do accounting of total usage to each record details is quite simple but massive in volume. No need for humans
I recommend trying Claude Cowork or Claude Agents, and then reassess what you wrote
@ab A glorified spell-check with a good PR firm.
It’s an amazing grammar tool. That’s about all it has going for it.
yes
This will make the rich richer. Unemployment skyrockets
@a6 but but our supreme leader predator said that
VZ AI stack would be ready really soon now!
“We are going to be substantially complete with that entire AI tech stack by July, and we hope to be fully done by November.”
If people can potentially make money from hyping it then everyone should be wary. I’m yet to see “it” do anything game changing in Verizon
No
Yes