Hacker News new | past | comments | ask | show | jobs | submit | AstralStorm's comments login

Funny thing, early deployments of a few Garbage Producing Transformers, and the base will be drowned out by the mess created and even humans will have problems telling what is right, much less a machine. Ever increasing amount to work out what is sense and what is a convincing lie.

Dead Internet Theory is already in motion.


Until it isn't.

IBM was a major player until it wasn't. Time Warner likewise. Etc.

There is no such thing as too big to fail, only a different kind of failure.


Direct CO2 capture is thermodynamically unviable, and literally every plan and attempt involving it was highly expensive in energy.

What does thermodynamically unviable means? There's no thermodynamic laws specifically on carbon.

I haven't crunched the numbers on this, but CO2 air capture is incredibly inefficient, and it consumes energy which itself releases heat. So if the amount of CO2 you capture reduces heat by X degrees but you release Y degrees in energy consumption to achieve that where Y > X, that would be one interpretation of "thermodynamically unviable". It seems implausible to me but not impossible.

The effect of the heat released in the air is negligible compared to the greenhouse effect.

CO2 capture might be economically unviable but nothing physically prevents it from working.


> The effect of the heat released in the air is negligible compared to the greenhouse effect.

And the amount of CO2 gathered by direct air capture is also negligible, so you can't just hand wave a position that one negligible amount is obviously greater than another.


Are you familiar with Terraform Industries works?

Unfortunately the plan is very dangerous as tectonic activity has a tendency to release it plus any hurricane or monsoon or thaiphoon or such has a tendency to destroy the installation or worse, move it somewhere where it will do damage to the ecosystem.

Also, use solar and wind ships instead. We don't need to sink more nuclear material...


The moment Google gets infiltrated by rogue AI content it will cease to be as useful and you get to train it with more knowledge.

It's slowly getting there.


It's been infiltrated by rogue SEO content for at least a decade.


Maybe, but given how good Gemma is for a 2b model I think Google has hedged their bets nicely.


Training is rather resource intensive either in time, RAM or VRAM. So it takes rather top end hardware. For the moment, nVidia's stuff works best if cost is no object.

For running them, you want a GPU. The limitation is that the model fits in VRAM or the performance will be slow.

But if you don't care about speed, there's more options.


Really except for the networking (including say Bluetooth) nobody is big endian anymore. So how about just don't leak that thing from the network layer.

And do not define any data format to be big endian anymore. Deine it as little endian (do not leave it undefined) and everyone will be happy.


I think both SMB and 9p (Plan 9 resource sharing/file system protocol) are little endian.

So it's not even all networking... and "network byte order" will mess you up.


For the more scummier or illegal elements on the network, that is true. For onion services, lasering attacks and takeovers plus honeypot are the chief danger.


Or you can use more... underhanded means that never result in an arrest.


And if you BYOIP, and run a large node, Tor volunteers will try to contact you and verify...


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: