qgin

I love Claude's memory implementation, but I turned memory off in ChatGPT. I use ChatGPT for too many disparate things and it was weird when it was making associations across things that aren't actually associated in my life.

I think it depends on the application. Employees tend to not be free in the 21st century at least.

We're likely 5-10 years out from a world where tappy-tap computer keyboard jobs are in a death spiral and caregiving jobs are one of the only fields untouched by automation.

I don't understand how many other developed countries are able to avoid the supposedly unavoidable perils of social services, but we act like this is a wild experiment that has never been tried anywhere in the world aside from the USSR or something.

Just look at this distinguished gentleman

https://upload.wikimedia.org/wikipedia/commons/thumb/6/61/Fi...

AI doesn't have to replace human jobs as they exist today 1:1.

In many situations, the work is not indivisible. If AI can handle 80% of the work, then a company can let AI handle that 80%, fire 80% of their people, and consolidate the remaining 20% still-human-work with whoever is left.

If you can turn off whatever news you’re consuming and not ever notice any effect in your life or in the life of anyone you know that would make you think “wow, something must have happened”, then you’re just consuming that news as entertainment.

Am I taking crazy pills? I love Next.js. What React partial SSR solution is better?

Social shaming is a big way humans deal with unchangeable things. They impose a cost for anyone expressing a desire for that thing to be different.

And it makes sense, really. You can't have a functioning society if everyone is running around freaking out about death all the time.

But we're entering a weird time where we might actually be able to add more good years to our lives. One of the steps towards getting there is being a little more okay with people seriously exploring these ideas.

New research reveals horse-drawn wagon engineering gains slowing, travel at speeds beyond 12mph unlikely.

When you have general anesthesia, you experience nothing, not even the vague sense of time passing like when sleeping, it's just an edit in time. You basically don't exist when you're under general anesthesia.

If a little propofol in my brain can make me not exist, I'm pretty sure when I don't even have a brain I will definitely not exist.

Or is that message the first challenge of the test...

I still maintain that we're not just a simulation, we're a screensaver on the computer of some extradimensional office worker who has been out of office for their version of a weekend.

The reason Meetup worked was years of collecting contact info of people interested in specific topics and then blasting all of them when a new group started in a location. If 1% of people responded, that was still a decent turnout.

This is the big barrier to entry for anyone trying to replicate Meetup. It was never about the event organizing software (which really wasn’t that great) it was about connecting your group or event to enough interested, local people that you had a chance at actually getting off the ground.

Meetup was bought by Bending Spoons (after a winding path getting passed around after it being sold to WeWork). They’re basically just running it until the wheels fall off.

THANK YOU

I guess where I'm coming from is there seems to be a lot of effort and energy available for discouraging and even banning this sort of LLM use without remotely comparable amounts of effort and energy put into figuring out how everyone can get access to mental health care from humans.

I'd bet a lot of money that very soon we'll have LLMs that are guiding people toward outcomes as good as (or better than) human therapists. I'd also bet a lot of money that we'll never manage to actually provide access to human therapists to everyone who could use it.

The choice between human therapist and computer chat is not a choice that most people in the world have. Most humans do not have access to a human therapist.

We should absolutely be talking about how to make LLM systems better at handling critical situations like this. But those that suggest that people should only talk to human therapists about their problems are taking a very “let them eat cake” position.

The current MacBooks are the best they've ever made. That wasn't true of the last generation. Late 2010s models were extremely cursed.

Many people in the past would have said reasoning would be impossible based on the same objection.

Humanity has a pretty extensive track record of making that declaration wrongly.

We didn’t design these models to be able to do the majority of the stuff they do. Almost ALL of the their abilities are emergent. Mechanistic interpretability is only beginning to start to understand how these models do what they do. It’s much more a field of discovery than traditional engineering.

It might be reasonable to assume that models today have no internal subjective experience, but that may not always be the case and the line may not be obvious when it is ultimately crossed.

Given that humans have a truly abysmal track record for not acknowledging the suffering of anyone or anything we benefit from, I think it makes a lot of sense to start taking these steps now.

That's the capillary refill test which tests circulation and perfusion. Doesn't really tell you anything about oxygen levels.

No proof, it’s just incredibly convenient. Just like when Kristi Noam just happened to get her purse stolen by someone who was in the country illegally right when the ICE raids were about to start. In this case, the well-known DOGE intern just happens to get carjacked in the city limits of DC right when Trump’s new DC Attorney General is being installed and the National Guard is ready to go.

No proof, but wow do they just happen to get exactly the event they need for the PR.

Is this different than what GM has been doing with Ultium?

If your art can be replaced by a model that recycles what’s already been done, maybe you were just recycling what’s already been done too.

There were even people who felt this way about anesthesia too. To survive in that era, they had to believe that suffering was somehow good for healing (or at least good spiritually) and talked about anesthesia as if it were cheating the natural god-given order.

It’s going to be a great time to be an entrepreneur, but a terrible time to be an employee.

A solo indie developer can have an AI team working on a project.

But so will every corporation. The competition for the remaining human-only roles will be intense.

I think the idea is the end of the game is nearing (AGI) and specific dollar amounts mean less than the binary outcome of getting there first.