Well done to openai for scraping all available code from the internet and amalgamating bits of it produce an apparently working program. We should also give credit to be many thousands of real people who manually curate their LLMs day after day. Let us not forget that there is minimal intelligence in current LLMs. Their only intelligence is that they can interpolate between datapoints to give outputs that are somewhere in between them. They have no ability to operate outside of their training data and therefore cannot provide truly novel solutions. Until they can show evidence genuinely novel solutions, and there is so far no fruitful research on this, I'm not going to worry about losing my job to an LLM.

This! AI has amazing uses in pattern matching applications, like spotting cancers in X-rays better than humans (and much much more), but it feels like the real power of AI has been lost in the noise of "we can replace your workers" narrative. Unfortunately, top level managers are always on the lookout for "we can cut your costs by reducing your workforce", which is (IMO) the origin of the techbro hype that's coming from all directions.

I always thought foreign aid was more about power and influence and being nice.

Silicon valley tech giants seem to be engineering an AI tech bubble, using employees as cannon fodder to give the illusion that AI in its current form can replace employees. Take microsoft as an example, they commanded their software devs to develop 50% of their code using AI, then fired a load of them with the public announcement that their AI product is so good they can fire their own staff. If AI can take over software jobs (which I personally don't believe will be the case), I'd expect to see an initial increase in jobs, primarily requesting AI skills, then a gradual decrease in jobs as those positions get replaced by AI. What the silicon valley narrative appears to be is "Our AI tools can replace your empoyees NOW!!!, fire your employees and use the savings to buy our AI offering".

Yeah exactly! Come back when it can write quality unit and integration tests that can be used to prove that the new code works the same way as the old.

I'd bet money that they have a huge list of fallback options, all kept secret of course. As soon as one gets uncovered, just issue a pretend apology while switching to the next fallback in the list.

Another great answer. Thanks

Great answer. Thanks

How can a passive device convert a lower frequency and therefore lower energy light into a higher energy light? Where does the extra energy come from?

It's a classic pitch to the middle classes. Get them worried about something and they'll gladly buy the expensive solution. In UK you can get a dumb phone for £13.50 (18usd). https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.argos.co.uk%2Fproduc...

LLMs have very low "intelligence" and have no ability to truly create. All they can do is write code based on an amalgamation of github, stackoverflow etc. If the programming languages and software requirements never change then, yes, maybe we could keep churning out code from the same training data. But everyone knows that software is ever changing, which means we will still need an army of developers to write the new programs that can feed the training data for the LLMs. If he's suggesting that we can have AGI in the next five years, well that'll never happen with LLMs which essentially fit curves to datapoints. With regard to AGI, there has so far been no meaningful development in this area to date. The rate of improvements to LLMs has greatly decreased recently because they've trained them on pretty much all available training data! At least that's Francios Chollet's recent take on things (I'd highly recomment listening to his podcast with Sean Carroll's Mindscape)

population when I say Fuck right off google

Agreed. For developing countries AI will likely make it much harder to get out of the middle income trap https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fen.wikipedia.org%2Fwiki%2F...

I think this is more to do with economic factors than AI taking over. You can't just replace programming jobs with AI, because you still need someone who understands programming to command the AI what to do. Otherwise you'd end up with managers writing "make my software prodcut better than all the competitors". I accept that maybe some of the most menial jobs could be replaced by a highly skilled programmers who can just crank out all the boiler plate code through an AI, which might result in a dip in junior positions. But if you're an employer with X skilled programmers, churning out code at a rate of 2Y, using AI, why wouldn't a company choose to double their output, and therefore revenuer, rather than keep the same revenue, but with half the programmers?

My Pixel is 2 or 3 years old, with moderate usage and the battery is starting to get lame already. Whilst the older pixels weren't too difficult to get inside for battery replacement, I found the more recent ones impenetrable. So, ironically, if you want your android phone to last 8 years, don't pick a google device.