Meanwhile Google is murdering their websearch business with chatbots. Rightly so, as that business needs to die, or at least materially transform.

Elon Musk sold it to Elon Musk. The price must have been meaningful. Wait till he does it again!

Because it contains information. Content > form, this is one of the cornerstones of hacker culture actually.

I used to read forums for schizophrenics (self disorders fascinate me, look the term up if you want to understand schizophrenia), and it was the consensus there that out of all the recreational drugs cannabis caused the greatest deterioration in one's mental state. Those are generally fairly sick people, but I don't think one can just ignore this signal. I personally went catatonic once after consumption, and I'm not schizophrenic at all. And that was in Amsterdam, so it wasn't some trash spiced up by a 17-years old dealer with whatever he found in his grandma's medical cabinet.

I don’t know what things are like in Amsterdam, but weed being “high quality” is not a good thing. The potency has skyrocketed over the past few decades. It’s like saying you went to Amsterdam to drink alcohol for the first time and had a bottle of tequila and passed out and puked; the fact that it was high quality tequila does not mean it’s better than a low quality glass of wine or beer. The issue is the doses are way too high.

Continuing the analogy, it does mean that they can blame the alcohol and not, say, rohypnol for their outcomes.

> it was the consensus there that out of all the recreational drugs cannabis caused the greatest deterioration in one's mental state

Methamphetamine and PCP might take issue with this statement.

But imagine the kind of self-selection you'd find on such a forum. Furthermore, it's been long known that cannabis use can trigger schizophrenia in those who were already genetically predisposed to such. Your observation and the OP are not news worthy, imo. Bad outcomes have also been observed when combined with immature brains.

But for healthy adults, countless have used cannabis for generations without experiencing harmful reactions.

This one's my favorite: https://thereitis.org/mr-x-by-carl-sagan/

Weed can make you paranoid, every smoker knows this. It is patently obvious that it could trigger something worse, too, if you are already sensitive.

I think only young people in their weed honeymoon phase get defensive about this.

Acute toxicity from a variety of other intoxicants can, as well.

THC is a toxin!

I'd been a casual user for a while (very sporadically in illegal days, regularly once legalized). I happened to going through a fairly traumatic breakup, and felt that if it didn't actually help me get to sleep, it at least made the insomnia tolerable.

My life settled down, but I continued to smoke a tiny bit. One puff really, before bed pretty much every night. When things were crazy and awful, I didn't care about feeling mossy the next day, but as normal life resumed, it definitely started to affect my attention.

Then, the "intrusive thoughts" started. It's like having an edge lord in your head. Whatever you think about, the idea gets lensed(?) from the worst/most extreme possible viewpoint/conclusion. It's hard to describe, but it was very distracting, and often just depressing. And of course, waking up at 3am every night, this is what I had to look forward to. And sometimes, I'd take a little toke in the hopes I'd get back to sleep.

Here's the punchline: I quit smoking and it went away in two days and hasn't returned.

Problem with recreational marijuana is that it’s so insanely strong. It would be like giving a child 190 proof azeotropic grain alcohol and being shocked that they immediately vomit. I can’t smoke pot - it’s just too strong.

I’ll admit to feeling a bit dumber and foggier after a few weeks of ingesting cannabis nightly though. That’s a real thing.

I've heard that the reason why marijuana is so strong is because it was illegal. The sellers wanted to have stronger weed to make it easier to transport; much like how during prohibition, people would prefer to import distilled alcohol, instead of beer.

Same logic applies to explain the popularity of fentanyl.

It's weird that people argue it's better for you to consume extra burned plant material to get to the same level of high-ness. If it's stronger, people just use less.

Problem is that they dont, actually

Thank you. HN delivers.

This may be a good place to exchange some security ideas. I've configured my OpenClaw in a Proxmox VM, firewalled it off of my home network so that it can only talk to the open Internet, and don't store any credentials that aren't necessary. Pretty much only the needed API keys and Signal linked device credentials. The models that can run locally do run locally, for example Whisper for voice messages or embeddings models for semantic search.

I think the security worries are less about the particular sandbox or where it runs, and more about that if you give it access to your Telegram account, it can exfiltrate data and cause other issues. But if you never hand it access to anything, obviously it won't be able to do any damage, unless you instruct it to.

You wouldn't typically give it access to your own telegram account. You use the telegram bot API to make a bot and the claw gateway only listens to messages from your own account

That's a very different approach, and a bot user is very different from a regular Telegram account, it won't be nearly as "useful", at least in the way I thought openclaw was supposed to work.

For example, a bot account cannot initiate conversations, so everyone would need to first message the bot, doesn't that defeat the entire purpose of giving openclaw access to it then? I thought they were supposed to be your assistant and do outbound stuff too, not just react to incoming events?

Once a conversation with a user is established, telegram bots can bleep away at you. Mine pings me whenever it puts a PR up, and when it's done responding to code reviews etc.

Right, but again that's not actually outbound at all, what you're describing is only inbound. Again, I thought the whole point was that the agent could start acting autonomously to some degree, not allow outbound kind of defeats the entire purpose, doesn't it?

There's a lot of useful autonomous things that don't require unrestricted outbound communication, but agreed that the "safe" claw configuration probably falls quite a bit short of the popular perception of a full AI assistant at this point.

Huh? The bot can communicate with me freely as it sees fit. A "conversation" in telegram parlance is not time-limited, it's ongoing once established, so no it's not only inbound. It can awaken and ping me whenever it wants. This can also work if it's added to a group chat.

If you mean it's not outbound as in it can't message arbitrary random users out of nowhere, well yeah, and that's a very desirable trait.

I was worried about the security risk of running it on my infrastructure, so I made my own:

https://github.com/skorokithakis/stavrobot

At least I can run this whenever, and it's all entirely sandboxed, with an architecture that still means I get the features. I even have some security tradeoffs like "you can ask the bot to configure plugin secrets for convenience, or you can do it yourself so it can never see them".

You're not going to be able to prevent the bot from exfiltrating stuff, but at least you can make sure it can't mess with its permissions and give itself more privileges.

If you're really into optimizing:

You don't need to store any credentials at all (aside from your provider key, unless you want to mod pi).

Your claw also shouldn't be able to talk to the open internet, it should be on a VPN with a filtering proxy and a webhook relay.

Genuinely curious, what are you doing with OpenClaw that genuinely improves your life?

The security concerns are valid, I can get anyone running one of these agents on their email inbox to dump a bunch of privileged information with a single email..

Does anyone have an answer to the problem of an OS for a laptop? I'm thinking about strong security here, less so about privacy (which is doable, for example via a Linux distribution).

ChromeOS (most secure OS), MacOS (most secure firmware and still much more secure OS compared to non-ChromeOS competitors)

Because the patient is usually unable to handle such information correctly (the medical system sometimes too). And the whole-body-scan type of tests additionally pre-select for the high anxiety types.

Google has barely released a successful product in 20 years.

Depend on the definition of the "product". For example some banal cloud storage in which everyone competes. And it's an "old" product, despite being invisibly improved behind the scenes, just like at any other provider. Google has pretty competitive storage AND they are fully abusing Android integration for AND they have pretty good bundling of that storage with other products, including, you've guessed it - LLM Gemini. So say a person is not a professional user of LLMs like a developer burning tokens in a dozen accounts simultaneously. A person has a phone and eventually memory runs out, so he buys a one click Google storage for 4 bucks. And suddenly he has Gemini Pro included too. So why pay 20 bucks to Anthropic, when Google costs 1/5 of that AND has other stuff bundled too?

So maybe Google is lagging on truly new products (btw, does Gemini itself with its TPUs count as a new product? I'd say yes), but "old" products are entrenched enough to carry them and compete.

I get 2TB (which I use) and AI Studio for $20, that's the best deal out there for me.

In a world where OneDrive exists?

I've used OneDrive across 3 companies with no problems.

Which one would you say is the best?

I'm back to Dropbox.

Google is good at buying existing products and scaling them, which is exactly what they did with DeepMind.

Deepmind was their worst acquisition ever. It is a vanity project that burns cash.

I thought that the likes of Android, Google Docs, Google Translate, etc. were fairly successful. Chrome and ChromeOS also seem fairly popular too.

A lot of those are getting pretty close to 20 years ago.

This year:

chromeos is 17

android is 18

chrome is 18

google docs is 20

google translate is 20

In retrospect, it is wild how good/successful google was 17-20 years ago!

Few years ago, we had Google Bard, the ancestor of Gemini, which was supposed to be an AI LLM, and when you right-clicked the page, it was a fake page with hardcoded sentences in a .js file...

> was supposed to be an AI LLM

Are you claiming Bard wasn't an LLM???

None of them looks like their original form, none would survive without google’s enormous investment.

Google Cloud is good and successful. Except they can't implement billing hard caps, or pretend they can't.

Im not sure what you consider successful. They've been struggling to get market share vs azure, and the product isnt that good. lots of rough edges, and piss poor support

Their API business model seems to be hope enough people accidentally go over free tier: $0 for the first 5000 monthly places lookups, $40 per 1000 after that

Neither does AWS and you can argue they aren't good but they're objectively successful, so it doesn't seem like a good metric.

But this is a phase change process.

Also, the temptation to shitpost in this thread ...

I read TFA. They found a best fit to a hyperbola. Great. One more data point will break the fit. Because it's not modeling a process, it's assigning an arbitrary zero point. Bad model.