The technology is fascinating, and does have many use cases and places for great integration. But the way the executive suites and charlatans on social media drag it’s capabilities through the dirt and claim it is going to dominate human tasks… it’s so insufferable. It can spit out a bunch of boiler plate code, but at the same time, a dev could google and click through a few links to find their data with much higher quality and precision to the problem. The dev has the business problem and the company proprietary information in their brains when searching and developing. That sort of info could never be compartmentalized into a LLM. In fact, it more shows how impressive humans are. So, it saves a few clicks, at the expense of being a complete black box with no rhyme or reason as to “why” or “how”. Even worse, the outputs it gives are based on training and human labeling from the bias of humans paid by a corporation…. That doesn’t exactly give me warm and fuzzies about the filtering process, quite the opposite. I just can’t wait for the next crypto bull run so all these 💩 heads can fake it somewhere else. At least it can kill no-code and save us all that headache.
Agreed on the insufferable part, but the business problem and proprietary information in the human brain is simply a technical hurdle and will be solved in the next few years.
How? The LLM converges on gradients. And business problems are constantly evolving and changing, and half the time you have to challenge what the ask is or evolve it. LLM’s give info no matter what you ask. A dev can challenge the ask or help reshape it.
Backtalk is already built in.
Welcome to the tech spin cycle
This. And Google is such a unwarranted casualty from the GPT hype. They kind of flinched and pushed Bard, but should have held their cards close and let GPT/Microsoft humiliate themselves. Oh well. Google search >>>>> LLM
The point is, it’s a gold rush for prosperity in the future. Not now. 10yrs down the lane, the richest and most IP valuable company could be an AI company
We are in a very early stage, like someone discovered the oil while digging. We will hace oil purification into different types of fuels, and then we will see consumer products built with the same oil. Hype will be settled slowly, and when dust settles, you will see real benefits.
awkward cope
How is it a cope? I’ve deployed multiple LLMs, they’re 💩
look within
Why SWEs are always whining about LLMs. Let them be.
Because, some of us are being told to develop them and innovate with them and do ops for them. And they’re a technology not suited to be deployed for anything other than a tutoring app.
Look up the human labeling factor for ChatGPT. They contracted humans to label data for training. If a human labels something there will be an inherent bias from said human.
In some sense I like the overly native ChatGPT hype. Especially that hysteria all over the Internet that "LLMs are gonna replace software engineers". That with the current badly managed mass tech layoffs will scare people from entering tech careers. The hype will die off. The hiring will recover. But the damage is already done. Demand for engineers will exceed the supply a lot in the next several years. Which means higher salaries and better job security. Basically, MS and Google etc are shooting themselves and the industry in the foot.
🎯
Chatgpt isn't even a year old. It's already made a huge impact people's every day lives.
What impact? Can you say it actually is better than a google search? Or, God forbid, having to think and write an email yourself?
I’m glad bears like you exist because they’re like people who discourage applying for a job. More opportunity on the table for me as a power user or candidate, thankies.
I like how charlatans like you hype these things up. The more layoffs now the more demand later. Enjoy.
2024 Presidential Election
Yesterday
597
Heartwarming peaceful protests
India
Yesterday
541
Modi is a legend, will be remembered for centuries to come
Tech Industry
Yesterday
3719
Asians - what are your thoughts on asian female white male ?
Tech Industry
Yesterday
1292
The end of Backdoor Roth?!
India
Yesterday
617
Who are these retards asking for dictatorship in India?
We already make money using an LLM feature in prod. You need to know where to look.
How?