Why count them at all?

It’s not a defense of the LLM. They are bad tools: LLMs can be unpredictable and if you hand them a nuke they may trigger it for whatever reason, even harmless and unrelated ones.

So don’t hand them a nuke, idiots.

I was kinda for it, but with the dangerous rise in quasi-nazi party AfD, I wonder what’s the point of banning a nazi flag… and it’s even more ridiculous that foreign ethno religious supremacists enjoy that level of protection.

Of course I’m not invested in individual transportation companies. Even less with the ones refusing to set on EVs.

Not mentioned in the article but Gemini is obviously developed by the DeepMind teams, among many other things.

I don’t think utopias are a bad idea in general, but if they somehow are only reachable by collapsing most of the current system before any groundwork can even be attempted… a form that can be developed in parallel and take over at some point makes much more sense.

Think about the transition and hopefully it doesn’t require nuclear war.

No, it has to do with growth. An adult eye no longer grows significantly.

Had a similar experience when SCUBA diving recently. New pressure gauges these days are digital and I still think the analog ones are not only prettier, but also functionally more convenient. You don’t need to be able to read numbers to know you’re getting into the red. Maybe they have some extra feature but I didn’t need it.

You don’t sell the same amount of product when you have to increase the price. You may need to shrink your business to not get the remaining margin getting eaten up by operational costs.

I agree that 0-days aren’t numbered. There are so many layers on which tech can be exploited that this is a difficult claim to make.

On the other hand, there are two different kind of exploits: clear holes in the logic, a situation or code path not considered by the coder. And the much harder to catch extremely creative ways to make a program do things it was never designed to do.

I have not seen LLMs doing creative things ever, so I doubt it would catch this second category. But sure, catching some logic holes it can be helpful with.

The job market is actually pretty bad right now and with all the recent layoffs in tech very saturated. Unionizing would make more sense.

But we have to identify this as what it is: an internal policy failure where they abandon proven processes to maintain code quality.

I guess I’m lucky my managers have not put that pressure on me yet. I do however see developers getting sloppy and lazier so the reviews actually do take more effort and AI rarely catches all problems with a change.

You wouldn’t want many carparks in the first place.

At least in my experience these models are pretty good now to write code based on best practices. If you ask for impractical things they will start doing ugly shortcuts or workarounds. A good eye catches these and you either rerun with a refined prompt, fix your own design or just keep telling it how you want to have it fixed.

You still gotta know how good code looks like to write it, but the models can help a lot.

Or rather the right to use shovels under ToS that can be changed on a whim.

I really don’t get this quantity-first approach. If you wanted to actually transform the world with tech in a way it’s not just superficial, you’d create task forces that sit together with specialist in each field of medical, construction, logistics, finance etc. give them 2 years to build prototypes and action plans. Then bet on the N most promising applications, spin them off as separate companies with premium access to your most advanced AI models and vertically integrate them into their workflows.

This would actually, sustainably achieve a foothold into these industries, disrupt and transform them long term.

He’s majority shareholder or has some trick to never be dethroned.

I have acquaintances at Meta and they literally waste tokens on bullshit tasks. They have like 10 agents running simultaneously doing some elaborate task that takes a long time. You can’t tell me this is more productive or efficient than doing actual work. Even if half of these tasks are somewhat useful and related to your project.

If I said the same thing about Judaism I’d get arrested in Germany.

It sounds like they were measuring chatbot use than a deeper integration into their systems. That may not me the best use of LLMs.