And when you say it like that, I have to wonder how much of this is a natural consequence of RHLF on such a grand scale, when you have millions of people pretty much much skimming chat responses or operating outside their depth and giving unqualified feedback to the models.
Seems like a lot of people may be reinforcing what sounds smart over what is smart.
Also as an aside: funny how much the LLMs continue to mirror the human communication they’re trained on
I believe we are several generations past peak-RLHF at this point. Now it's much more RLVR (Reinforcement Learning with Verifiable Rewards), with a goal/evaluator loop.
Which, conveniently, fits neatly into the benchmaxxing arms race/agentic coding market fit, since you can basically train "directly" on a specific problem space for a benchmark/agentic goal (fudged sufficiently to avoid excess overfitting on public problems/bechmaxxing accusations if real world performance falls short).
The language evolution could be explained by reliance on ever increasing layers of a model judging a model, using a model developed eval, based on synthetic data from a model, etc. And by the time a human evaluator sees it both A/B choices already converged into weird Claude pseudo English as that was baked in much earlier in training.
This, 100%. I don’t think the industry knows how to scale LLMs’ general intelligence much further. The training paradigm is about maximizing very specific behaviors / very specific tasks, but doing lots and lots of them. Which can create the illusion of general intelligence if your tasks are similar to the ones the models were fitted for.
> or operating outside their depth and giving unqualified feedback to the models
I wonder if the labs are sufficiently prepared to filter this kind of stuff out. I see a lot of non-developers asking development things of Claude, getting confused when they're in over their depth, and getting upset that they don't understand what the model is providing them, giving it bad feedback, and subsequently making the AI worse for the rest of us who know how to use the tool.
The “whining” stems from watching the communication style obviously degrade, and it’s a huge problem for people who want to use this stuff to build and instead continually fight the tools.
Like so many other products, people are moving too fast and shipping things that move the ground under people’s feet needlessly.
All this while we’re beaten to death with the marketing and false promises, and the broader consequences (ex: layoffs, stress, crazy expectations) caused from all this.
Obviously what Anthropic and co have built is amazing and people aren’t losing sight of that. That’s actually the key part of the frustration.
So no, this is not whining. This is the natural response you get when you make bad product decisions.
If you don’t want to get feedback, don’t sell products.
Can you expand on why you chose to use CRDB with Oban? I have no opinion here, I’m genuinely curious as someone using Oban myself (with Postgres). I haven’t hit the point of really needing to scale it out yet and I’d rather avoid the traps others have figured out.
Ah ok thanks for the clarification. And thank you sorentwo for your fantastic work – I've been loving my switch to the Elixir ecosystem thanks to the efforts of folks like you.
> (because, you know, I was getting paid to write code and that's the only way I could actually get it done)
I'm going to assume you were getting paid to build software that solved problems and created value for your customers and stakeholders. Writing code has always been just one activity that's part of the job, and developers forget that and make statements like this! That's the parent poster's point. I'm not saying it's not an extremely important part of the job, or that people don't often collaborate poorly in ways that take away from the sacred deep work time, but framing it as "I get paid to do X and not Y" is just a highly limiting way to look at or talk about the role.
>> (because, you know, I was getting paid to write code and that's the only way I could actually get it done)
> I'm going to assume you were getting paid to build software that solved problems and created value for your customers and stakeholders.
That's a distinction without a difference. At least historically, I was "paid to build software that solved problems" and I was to do that by writing code. If I didn't write code, and enough of it, I'd be fired. Getting my flow state disrupted for no good reason was something I'd resist.
Also agile ceremonies are a drag, literally becoming the thing agile was originally supposed to be fighting against (not that agile is gospel, I've always disagreed with some of its practices). They're not a good reason. And I also mentioned an actual good reason. I should also note those meetings I was referring to positively were almost always with users, not tech people.
> Writing code has always been just one activity that's part of the job, and developers forget that and make statements like this! That's the parent poster's point.
I wasn't addressing the parent poster's point per se (and I noted that and why), just noting that a lot of the "collaborative" activities he cited were often not that collaborative, and the shade he was throwing at people who were unenthusiastic about participating in them was probably unwarranted and misguided.
tl;dr: OP needs to have more empathy. There are better ways to thread the needle of his observations than what was on display in his comment.
I've worked with engineers all over the spectrum in terms of their styles, beliefs, and preferences... and some of them are frankly not very interested in getting out of their comfort zone (like heads down, writing code and being alone), and optimizing for the group rather than themselves.
So yes, they are in fact unhappy to communicate (in a general sense), because of how tedious and uncomfortable communication often is.
I'm not saying it's irrational or immoral, or not driven by the types of past poor experiences you mention, but in my experience it's often pretty obviously suboptimal and highly frustrating to work with.
Some are inherently that way. Yes. Those don't thrive even within the engineer groups. Outliers don't make the rule. OP argues developers are hypocrites for demanding to be left alone when it suits them, or engaged when they feel they job is at stake.
If anything, genAI for coding is making engineers seek more engagement, that can't be a bad thing given the fracture between the business and product development.
How do you think it changed the world? I don’t think that was an especially prescient thing to say/write at that time. The idea that software was poised to continue to grow in 2011 was pretty obvious to most people. It is true that some companies were undervalued and many VCs and other folks were scarred from the dotcom bust.
But if you go back and read it, you might notice that a lot of the companies and software he discussed and predictions along with them failed to be true or lasting.
They’ve lost a whole lot of people in prominent roles over the past few years. I wonder how much of the misfires and general thrash in product direction is a result of brain drain and/or so many hands changing. Or maybe I’m confusing cause and effect… hard to tell
That’s pretty crazy, I swear it wasn’t that long ago these companies were about the only people hiring and the comp packages looked absolutely deranged.
For a brief moment I regretted wasting any time of my life on anything but ML research. But I guess the bigger they come…
This project reminds me of a book I highly recommend called An American Sickness. It sheds a lot of light on the same sorts of issues.
One underlying, perverse incentive behind many of the problems is that insurers are regulated based on percentages of spending rather than total costs.
The US passed laws meant to limit marketing and overhead that tied insurers economics to the size of the overall medical bill... which means as healthcare spending rises, the dollars they’re allowed to retain can rise too, which basically means they're incentivized to drive costs up rather than down.
> The US passed laws meant to limit marketing and overhead that tied insurers economics to the size of the overall medical bill... which means as healthcare spending rises, the dollars they’re allowed to retain can rise too, which basically means they're incentivized to drive costs up rather than down.
Yes, this is an important piece of the puzzle. The "medical loss ratio" for large insurers (the kind we all know and love) is set to 85%. So they can keep up to 15% of their revenue as profit.
As you said, if total spending goes up, they get 15% of a larger number.
It's almost as if the insurance companies wrote those regulations. The same ones that required everyone to purchase their product and implemented government subsidies to pay them. Legitimately no way anything other than price increases and insurance profits could happen.
I originally thought AI-assisted writing would help synthesize what felt like original ideas I had, that I just wanted to get out there without the laborious task of editing. I didn't expect the writing to end up feeling so incredibly tired and watered-down, but upon more reflection on how the models actually work, it's not all that surprising. Uniqueness in writing is both in the style/structure and the message, and all AI seems to do is find the local maximum of both. Lately I've found myself going back to writing things myself (not all the time, depends on the task), and wishing there was a way I could just completely eliminate the slop from certain things I look at. I worry about all our minds, and the garbage-in, garbage-out net effect of this.
Maybe that's why the writing feels so terrible. The AI is attempting to maximize every sentence while simultaneously expanding on just a few actually meaningful points. And the net result of that dissonance is this rage-inducing vapidity. It's the written equivalent of the Uncanny Valley.
And when you say it like that, I have to wonder how much of this is a natural consequence of RHLF on such a grand scale, when you have millions of people pretty much much skimming chat responses or operating outside their depth and giving unqualified feedback to the models.
Seems like a lot of people may be reinforcing what sounds smart over what is smart.
Also as an aside: funny how much the LLMs continue to mirror the human communication they’re trained on