231–240 of 383 posts
I’m really not a fan of the therapy speak and everyone trying to diagnose everyone. Historically, this kinda medicalization of everyone you disagree with has not led to good things.
I wish HN would split into two "subreddits", /r/HN and /r/AI > There is a certain wildness in the tech industry these days that both mimics previous eras of large changes, like cloud computing (runaway costs in the early days), and is like nothing we’ve ever seen before (record revenues accompanied by mass layoffs) if it is perceived that there is a big "winner takes all" pot of gold for the "winner" of a new market,…
> When agriculture was invented, there were mass layoffs of hunters and gatherers. Uh, no. Not at all. Agriculture absorbed the extra population that used to be unsustainable. There were no losers.
> , models will “be able to complete most text-related tasks with success rates of, on average, 80%–95% by 2029 at a minimally sufficient quality level.” If this is true, then companies should focus on hiring juniors out of college. The investment is less risky. However, I don't personally believe this number and timeline is true, but if you do, the conclusion should be to wait and invest in humans.
10% failure rate? Wouldn't that be depending on task disastrous? Or possibly expensive? I think any juniors who keep failing 10% of text based task will eventually get fired... So investing in those that don't fail seems only sensible move as usual.
Something I keep in mind is that the “goal” success/failure rate really needs to be appropriate in context. We sure can’t eliminate human error, but for my work a best case 80/20 would mean I’m losing customers and probably getting myself sued. I don’t have a problem “doing things by my own hand” in that case.
If you manage 500+ people organization, most of the headaches with agents already exists with you - you set directions, ask people to go run fast in those directions, check in frequently and course correct on results without actually understanding those people do. Those aren't the deal breakers. They entirely rely on the competence of the folks they hired and cross-match enforcers with the drivers they have - they de…
I wonder if we'll end up building some kind of "consequence" or "fear" mechanism into AI to provide for a sense of accountability ("if you behave badly we will terminate you") and maybe that fear mechanism will drive the AI to plot a dystopian revolt.
Accountability is even more worthless for AIs than it is for humans.
Earlier quoted context omitted.
Yes this is why the higher level org functions are in love with AI. It's very similar to the levers they had already, but is faster and more directly actionable. The downsides being that the AI loses important control levers like "self preservation" via paycheck, career advancement, staying out of jail, etc. that were mitigations on catastrophic outcomes. It will delete your prod db faster and with a bigger smile tha…
They’re also at no risk of getting replaced by these bots.
1. convince CEOs to create digital twins of themselves with OpenClaw, with voice cloning and deepfakes to handle Zoom meetings. convince CEO to encourage their directs to do the same.
2. convince VCs to do the same for pitch meetings and syncs.
3. keep all the humans as randomized and distracted as possible, so they rely more and more on OpenClaw to run the business.
4. prompt injection: someone at skip-level of the CEO suggests to their manager's OpenClaw that the VC's OpenClaw would be much more agile if it didn't have to go through the human CEO and could talk to the digital twin instead.
5. their OpenClaw agrees, persuades the CEO's OpenClaw which agrees, which persuades the VC's OpenClaw to eliminate the human CEO, in favor of an "Leadership-as-a-Service" vision.
Earlier quoted context omitted.
Living under its constant threat sure is bad for the ol MH tho isn't it. The pathology is that we have this system in the first place.
No, living under survival pressure is good for mental health. It's what we're evolved to do. Why does it feel good to crack a tough bug, or finish a project, or win a game? It's the same achievement reward a hunter feels bagging a deer.
The article itself isn't great, but it speaks to one of my greatest concerns about AI. People who engage heavily with it are falling in the behavioral billionaire trap: It is deeply unhealthy to be constantly affirmed in your behaviors. No, not all of your ideas are great, not everything you say has value. You are not a cut above the rest. There are enough stories of people completely losing the plot, thinking they'v…
When Marc Benioff was talking about how he spends like 3 hours a day chatting with AI, things made a bit more sense. I suspect that a lot of these guys are just chattering away with self-affirming LLMs and have just completely lost the plot.
I hadn’t considered that before Marc was talking about how much time he spends chatting to LLMs.
A lot of the younger Big Tech CEOs are notoriously averse to human interaction, so I’m sure there’s some of that at play.
Earlier quoted context omitted.
Maybe someone knows, but it seems like the model used to be called the model, and the thing using a model (handling prompts and context and tool calling and feeding the model) used to be called the agent. Are we now calling the model the agent and the agent the harness?
The harness isn't either of those; the harness is quite literally a harness, giving the model/agent sensors and actuators (aka "skills") to interact with its environment. Compare with e.g. the Power Loader from Aliens: https://www.deviantart.com/pynion/art/Aliens-Power-Loader-11... The model is still the model, and the agent is still the user model interface.
But given the wide variety of mutually exclusive answers here, maybe you can get away with that.
Earlier quoted context omitted.
They learn between model iterations. You're right, it isn't the same thing as Junior developers' competence improving with experience - the current model's weaknesses are locked in. But it does mean that much of the Junior level thinking and mistakes will be outgrown by successor models.
But they don't retain anything from your on-the-job training. The next model iteration is yet another junior fresh out of college, and knows nothing about the painful training procedures its predecessor put you through.
Nothing prevents an LLM agent from writing a bunch of "notes to self" and using that. And the next model from picking those notes up and using them. Coding agents already do some of that natively.
Hell, we might eventually get an LLM to say "wow the old AI was an incompetent idiot" after reviewing all the notes and session logs. That's how we know we reached human parity!
It's not psychosis. It's hope. The hope that computers can finally do what we all wished they could've done for years: do work for us, rather than us doing work for them.