This is such a tired, meaningless argument. I've never seen a human in 10 years of professional software engineering at a large company ever so confidently, consistently create and send out seemingly well-reasoned code that's as wrong as what SOTA models using CC or Codex do. If a human did this, they would be fired or perpetually remain a junior who no one wants to work with.
Also, if a human does this, you can replace them and get a human who will not do it. The default for an LLM is to generate plausible-looking text that may or may not be completely incoherent. That is not the default for a human. Again, if you find that your colleague consistently fabricates APIs, you can hire someone who isn't crazy instead, but you cannot do the same with LLMs.
This is commonly known as "LLM-as-a-judge" and anecdotally multiple people I know who write code using OpenRouter or using multiple models say it's surprisingly effective. It's strange that there don't appear to be any major papers on it since ~early 2025, which at this point is basically ancient history.
Ah yes, the magical equivalent of "you are a senior software engineer who writes bug-free code".
IME people would benefit greatly from the process, albeit tedious and time-consuming, of testing out the same prompt sequence/session with the exact same model multiple times. It becomes clear extremely quickly how capable but unreliable and inconsistent a model can be even when given the same context. If you have ever completed a long, complicated task with an agent and then lost the session and tried doing the same thing again from scratch you may have had the experience of seeing the subtle changes that come up in the model's thinking which lead it to accept or reject certain paths and ignore or incorporate prompt instructions like the one you've provided.
Do you know if anyone has trained, say, a pre-2017 model and tried to get it to come up with Attention Is All You Need? If it did, would you say that was only because it's a synthesis of prior art? If so, what isn't?
Allow me to restate my point: human beings and AI both create via synthesis, but we are the only ones capable of what we could categorize as true original thought or creativity. It could be argued that nothing we do as humans is truly original or creative either, but I would counter that with the claim that an LLM could not have created any element of the society and culture that gave birth to LLMs. Maybe in six more months.
Not OP, but in 6 months of using Opus I haven't yet found anything that I know how to do but it does not. On the contrary -- it can do things instantly that I would have needed a ~week refresher on some SDK or some algorithm in order to implement myself--plus a ton of thrash/debugging time.
like, you want to watch an AI play a video game and throw a baseball? that's like saying my toaster can't solve math problems what a shitty toaster amirite
Cool opinion pieces that build stories from cherry-picked data.
My sources? Easy.
Just try searching for news stories of companies directly stating AI is the reason for layoffs. I don’t need to do any statistical or mental gymnastics for my evidence of it happening when it is literally as easy as looking at the result and the stated causes.
Your first article seems to start with a premise that laying off workers, whose jobs magically no longer need doing, and hiring AI focused workers means AI didn’t replace anyone. Your second article really doesn’t seem to address the subject beyond mostly talking about “overhyped” versions of conversations. The third article uses the fact that people who are exposed to AI aren’t entirely unemployed (despite rising unemployment across sectors and decreasing pay) and as such is proof a job apocalypse is unlikely. You follow up with yet another article in a similar vein.
Are you missing the pattern here?
Which of these articles and which part specifically do you believe supports the statement “NO ONE” is losing their job due to AI? Zero. Not a single one?
Okay, if not that, are you trying to refute my statement that automation is increasing across sectors? Which one of your articles refutes this statement categorically? Is it that you’re refuting AI is used in or to implement automations? Are you refuting that people have lost jobs to automations? Are you refuting that an AI does not need to replace an ENTIRE job in order to cause someone to lose their job or for a job to be replaced with a lesser paying job?
I’m not clear on how you think what you have provided is actually contextually relevant to the words I actually wrote, much less how I must have a “problem” because of it.
if you believe what companies are telling you (publicly traded companies especially which have to say this), I got some Enron stock options to sell to you :)
Not to be cynical but do you think this would matter at all? Are you saying that companies would hold themselves to their missions or even something that's legally binding?
> "Google is not a conventional company. We do not intend to become one."
> OpenAI being founded as a nonprofit and becoming for profit.
> Didn't Anthropic literally say they wouldn't train on your data or keep it for longer than 30 days unless legally required, and then decided to opt people in to having their conversations used for training?
if it's in the charter/articles of incorporation/ articles of organization, it's binding. If I break the mission and a.
> OpenAI being founded as a nonprofit and becoming for profit.
I think this is a common misconception, or a disregard for nuances. The NFP was not and cannot be converted to a Corp, that's kind of the idea of an NFP. However there exist satellite companies.
Sam Altman does not own shares of Open AI because there are no shares.
OpenAI has a for profit company (capped, Public benefit corporation), which Sama I don't think has shares in. It's an instrument for investments.
But every transaction needs to be fair and in kind, there can be no gifts at any point in a way that would magically negate the purpose of an NFP, Sama cannot cede the IP of ChatGPT to himself or one of his companies, that's not what's going on.
> Didn't Anthropic literally say they wouldn't train on your data or keep it for longer than 30 days unless legally required, and then decided to opt people in to having their conversations used for training?
Again, saying it, putting in terms of contracts (that can be retracted with notice), and putting it in the charter are all different.
Also, if a human does this, you can replace them and get a human who will not do it. The default for an LLM is to generate plausible-looking text that may or may not be completely incoherent. That is not the default for a human. Again, if you find that your colleague consistently fabricates APIs, you can hire someone who isn't crazy instead, but you cannot do the same with LLMs.
reply