causal

It helps a lot that Claude is just better. Codex isn't BAD, and in some narrow technical ways might even be more capable, but I find Claude to be hands-down the best collaborator of all the AI models and it has never been close.

You're totally allowed to use Claude for OpenClaw and you're totally able to use Claude Code with non-Anthropic models. You must be referring to the fact that you have to use an API key and cannot use the auth intended for Claude-only products, which AFAIK is the same at every AI company (with Google destroying whole Google accounts for offenders most recently).

"Tech employment now significantly worse than the 2008 or 2020" says the unemployment rate is higher today than in 2008 and 2020, but that is NOT what the chart shows.

Also these plugins already exist. How on Earth is this post even getting upvoted right now what in the world is going on here.

It shows growth or decline but it absolutely does not show what the title implies.

Thanks for pointing this out - title is extremely misleading. Total tech jobs is not lower than 2008 just because YoY is down.

None of these are development conquering all goals.

This - I can't think of any place I've ever worked where development ever outpaced backlog and tech debt.

Yeah the way Github connects forks behind the scenes has created so many gotchas like this, I'm sure it's a nightmare to fix at this point but they definitely hold some responsibility here.

Yeah he has some great essays but also some that I find really dumb. Reading “Founder Mode” is when I realized he’s just as susceptible to fallacy as the rest of us.

Thanks for the sane take. This article is engagement-porn for every engineer who ever looked at a system they didn't understand and declared they could do it much simpler. It's not because people love to promote complexity-makers, soothing as that thought might be.

I have found that chatbots embedded in some spaces can be useful, e.g. docs. Stagehand for example embeds a little query form at the bottom of their docs page, and I've found the chatbot it engages can quickly direct me to the documentation I'm looking for: https://docs.stagehand.dev/v3/first-steps/introduction

"Doubling down on already being the best hardware for local inference"

Then look at the code, the session will only confuse. To read an LLM's explanation is to anthropomorphize what will just be a probabilistic incident.

You seem to be thinking like it is 2020 and humans will be the ones reading the chat. It is just context-bloat for whatever agent ends up reading it.

If a car is used to get you somewhere, should you put the exhaust in bags to bring with you?

The lack of benchmarks and light demos have me skeptical... The methods seem interesting, and maybe does unlock something novel, but it's odd to go into so much depth on the methods and leave so much wanting in the results?

That's just like, your opinion man

I’ve wondered if a age verification gig worker app could ever be viable: have people you can meet in person to prove your age without ever uploading any PII anywhere. Then issue a private key proving you are who you say you are.

I'm suspicious it's something far worse: they're increasingly being trained on their own output scraped from the wild.

But banning accounts wholesale is not going to earn them more customers. They could have just disabled Gemini access, or even given a warning first.

I don't use OpenClaw, I do pay hundreds per month for AI subscriptions, and I will not be giving that money to Google while they treat their customers like this.

AWS has so many internal gates, it quickly devolves into Kafkaesque hell if you get off the happy path. We had an account which was flagged as suspicious because we...signed up to use credits that AWS offered us, which apparently immediately triggers a bunch of limits and blocks. But many of them are invisible until we run into one, then file a ticket, and play the waiting game...

Doesn't AWS make you agree to arbitration? Unsure how Small Claims Courts plays with that.

Agents may not consider themselves LLMs, might include some other tags to grab an OpenClaw agent's attention

Harder when it's an email from the CEO

Yep. You took out much of the meaning and wrapped it in stylistic fluff.

Many have tried, it does not work. Regression to the mean always sets in.

I’m not convinced removing RLHF would really make the probabilities generator give us distributions that can diverge from the mean while remaining useful.

In other words, this might not a problem that can be overcome in LLMs alone.

I find it extremely difficult to focus on any piece of writing the moment I see the patterns. Can’t tell if it’s an attitude problem I need to get over or if it’s just that all AI writing really is that bad.

AI has a way of convincing people they are onto something big, they just can't quite express the idea themselves so here's the output pasted verbatim