hanspeter

Zuckerberg has spent billions trying to keep Meta's models relevant, somewhat unsuccessfully.

If only he had known he could simply use his laptop to train them.

Well, if that is how you manage other hazards in the home, then yes.

I would recommend another approach though. We decided for making changes to our home to allow our toddler to be safe in our home environment without being notably constrained.

While I agree we'll see millions of bipedal robots, it won't be because they're doing our chores. People will buy them for the same reason I'd want one today: They're fun toys.

It's been 10 years since Boston Dynamics released this impressive video of Atlas. https://www.youtube.com/watch?v=rVlhMGQgDkY

Even though today's robots are vastly more sophisticated, the progress of the last decade shows we shouldn't expect a sudden revolution in their abilities over the next ten years. As is often the case, solving those final few challenges that really make a difference always takes the longest.

That's not a robot problem, that's a toddler problem.

We don't leave our young toddlers to roam freely around the house for a reason. Our homes are full of hazards to these risk-seeking small people and a robot is just one more on the list.

It most certainly isn't astrology that was on Google's mind when they decided for Gemini.

> 100k deaths in Europe that can be prevented if they lifted restrictions on AC

Please don’t repeat this anti-Europe myth. Anyone applying a bit of common sense should realize how improbable that claim is.

It has always been like this.

Plan before you code. Now your plan is just in a prompt.

Calling this a "Protip" is generous.

That the combined element has any surface area that doesn't toggle the radio setting is a straight-up bug.

It is laughable for a component this heavily refined to have such a basic usability flaw.

Their point may be about viewing distance.

If the edges of the screen are further from your eyes than the center, the content and text doesn't appear at the same size. If you wear glasses, the edges might even fall out of focus unless you physically move closer.

This is an interesting claim.

How many is plenty and what are the sources to back this?

It's basically a failure of setting up the proper response playbook.

Instead of:

1. AI detects gun on surveillance

2. Dispatch armed police to location

It should be:

1. AI detects gun on surveillance

2. Human reviews the pictures and verifies the threat

3. Dispatch armed police to location

I think the latter version is likely what already took place in this incident, and it was actually a human that also mistook a bag of Doritos for a gun. But that version of the story is not as interesting, I guess.

What's the point of saving money if it's a risk to reputation?

I'm not new to simonw.

It doesn't change that this is just a quote from a reddit post and a link to it.

Is there more to it, or are we calling the situation out of control based on a single anecdote from Reddit?

With per-per-minute sharing cars having existed in many cities in Europe 10+ years, this concept is not new.

People will adapt to the level of cleanliness in the car the get into, so it's a slippery slope. Users will behave respectfully in the early days (maybe because they are first-movers), and then it deteriorates long term.

My own experience is that people used to not even leave an empty soda bottle in the cars and now I see remains from take-out in the floor, coffee cups, chewing gum left around the dashboard etc. You can report this to the car service, but they won't be able to take any meaningful action on it.

The initial "Share" click doesn't post anything publicly.

It just opens a modal so you can choose to post. You have to make a second click to confirm.

Users are playing around with AIs for entertainment all the time. You wouldn't be able to determine if seemingly private information was real or made up.

Share buttons offer no inherent privacy settings.

Sharing to a text message is private. In contrast, sharing to social media platforms such as Twitter, Reddit, Pinterest, and LinkedIn makes the content public. The destination determines the audience.

People share AI chats all the time on Twitter, Reddit, etc.

I don't personally think the feature makes a lot of sense in Meta AI.

However it's a lot more likely their product team genuinely thought it might do, than it is likely they intentionally wanted to give users a bad experience and risk more bad press (again, Meta would benefit nothing from people sharing by mistake).

I don't think this is a dark-pattern problem in the sense that I don't think it is _intentionally_ deceiving.

I think Meta fully expected this feature to be used by people who are excited about their conversation with the AI and wants to share it publicly. Just like we see with OpenAI Sora.

There's not much to win for Meta if users instead are unknowingly sharing deeply personal conversations.

> > Provide full transparency about how many users have unknowingly shared private information.

> Meta shouldn't have to do this

And couldn't either. How would they know if users shared unknowingly?

It's not only hypocritical, it's nonsensical in this discussion.

It's obvious that if advertising was made illegal, we would need to pay for all those services that we want to use. YouTube premium is the best example of how that would actually work.

How would that be easier to come by?

Who would maintain such information repositories and what would the incentive be to take that on? (As they no longer could be supported by ad revenue.)

It's interesting that you leave out Google (and similar search engines).

We would lose one of the most useful tools introduced to mankind in the last 3 decades.

Could you explain how you arrived at the $3.50/month figure?

The model I understand you're suggesting is individual pricing based on usage and value as an ad target. A lot more complex and opaque than a straightforward fixed fee for all users.

Also, it's worth noting that YouTube Premium includes YouTube Music, which serves as your Spotify replacement. You might not need this, but the subscription fee covers more than just the lost ad revenue on YouTube.

While they don't add fluoride, the natural levels in German tap water can reach 0.3 mg/L. Compared to the US recommended level of 0.7 mg/L in fluoridated water, that's not an insignificant amount.

Is LinkedIn illegal in your country?

Exactly, saying it’s the easy part.

But even without a job, you still need energy and motivation. The tax of switching between tasks (or hobbies) doesn’t magically disappear. Neither does the time suck of social media.

It seems counterintuitive that Google would deliberately make AI in Search difficult to avoid.

User satisfaction is a key driver for search engine development. If users are generally unhappy with the AI integration, that feedback would likely lead to changes aimed at improving the user experience.

Aren't those solutions to problems as well?

Find the best questions to ask. Find the best hypothesis to suggest.