fidotron

No, that's other people being concerned, it's not the people in the relationships that are concerned.

The position was "a machine pretending to be a human can’t engender the same emotional response as another human" and quite clearly to some people it can, so the statement is false.

The problem isn't "Is this human or AI?" it's "Do we trust the entity with this identity or not?" meaning it's entirely possible to be anonymous but use a web-of-trust style system to promote/demote stories.

Part of the problem with that suggestion is everyone likes to tell everyone else who it is they should and shouldn't be trusting, so that dataset and acting on it would become a huge hot potato.

a machine pretending to be a human can’t engender the same emotional response as another human.

If this were so obviously true there wouldn't be nearly so much concern about people having relationships with AIs.

Most people likely have a natural aversion against having "team members" without agency or consciousness. Giving them artificial, acted personalities would be even more horrible.

The problem isn't the team member being artificial or not, it's the obvious fakery of the personalities that is the problem, and that's entirely independent from the artificiality of the team member.

Demonstration of personal responsibility and accountability?

Or is that too much?

This is an interesting step in a fairly obvious direction.

I wonder how they plan to make mobile work, and it's amusing going with an xyz domain given their reputation.

Technical staff are easiest to pin the blame on in any given situation. All those other fields require enormously more people skills and political inclination.

If you are in an org like that and there is an effort to pin something on technical people inappropriately the reaction when there's solid push back approaches total hysterical breakdown since it's so unexpected. Even if the tech staff "win" they'll be marked for the rest of their time there.

AI has forced a dramatic shift in my view on intellectual property. The laws as they exist only protect corporations now.

Even this isn't quite right. Publishers unambiguously have been screwed.

What has happened is if you can give the political/investment classes enough upside opportunity in the entities doing the IP theft then you're golden, and they increasingly have no problem with even pretending to hide it.

Yeah. People should absolutely be _trying_ the Chinese models, and experimenting with running things locally, but the noise in development is genuinely all Claude and Codex.

I put my foot in the mobile comparison the other day, and will again. If you were to go back and be a mobile dev in 2010 by all means specialize on one platform, but play with both as a professional interest to stay realistic. Here it's important people have access to US/Chinese/Other, open/closed, local/cloud and that this remains. Don't become a blind Claude guy or a open weights fanatic: that way lies disappointment.

The big decelerationist threat is a sudden reduction in competition. If either OpenAI or Anthropic drop out or the open weights stuff is banned/becomes uncompetitive then the motivation and tolerance for taking risks with the larger training runs tanks.

The closest we've seen to this in tech in recent decades was iOS vs Android, where Android only really was competitive for a very short window of time (approx 4.x) and it was during that period that both Android and iOS actually improved dramatically for end users. Once Android lost the plot again, and especially in the US market, all that energy started going in some very silly directions.

That link was fascinating, thanks! (Arguably worth a separate submission imo).

If you ask an LLM, "what known example did you use to solve this?" it is very likely to cite something plausible-sounding. That absolutely doesn't mean it is what it actually did. It is trying to give the "right answer" and please.

That's not what I said happened though. It didn't solve the problem (for weeks) until I (accidentally) told it which example it happened to know was relevant, and then it solved it in hours.

Aha, thanks!

Do you have any insight into what algorithm it uses? Like a ghost runner of the record pace or something?

"Lack" isn't the right word. "Lacking" is more like it.

Yeah, that's fair.

My takeaway is that text is a far richer modality than anyone has expected - and that high end LLMs are often sharp and flexible enough to recognize their weak points and substitute their strengths. I.e. all the LLMs implementing A* to optimally solve pathfinding in ARC-AGI-3 tasks, often unprompted.

I agree and disagree with this. I think we've learned a lot of humans are more text based than we thought, but conversely I'm not persuaded what non-textual task reasoning LLMs are doing is necessarily text based, just that models have grown large enough for other reasoning modes to conceivably be hiding in the parameter space.

As I mentioned elsewhere, like many others I find LLMs work entirely by example, and reaching for A* when pathfinding is the single obvious thing to do. In cases where the magic key word is not mentioned and the problem cannot be identified as "pathfinding" (or some other trigger with a highly specific widely documented solution) they will struggle, yet the moment the trigger is hit they get there very fast. This is why prompting remains such an art form.

Fable is the first one I've encountered that is capable of serious open ended 3D programming in ways that suggest it has some grasp of the spatial aspects of the problem (not merely symbolic manipulation of the vectors etc.), but it still misses optimization opportunities a human will find glaringly obvious based on spatially predictable bounds etc.

The mention of pacemakers made me wonder if you could have some light marker on the track to show the ideal pace you should be keeping as well.

If you can find the human equivalent of the rabbit for greyhounds then maybe even more could be achieved.

I will be posting something to that effect later this week. (Hopefully).

Basically current gen LLMs apparently do spatial reasoning the way they seemingly do everything else: by reference to previous example. I didn't see them work out which known example to use for a given problem until specifically prompted, in my case by accident.

And that's why you aren't qualified for a customer service role but might be for something that current AI is competitive with.

What a silly and obtuse comment.

You mean the answer betrays the point: customer service is surprisingly hard, we just have a large number of people that are capable of doing it.

This is what the whole https://people.csail.mit.edu/brooks/papers/elephants.pdf is about.

Do we employ mathematicians in customer service roles?

It's still clear that LLMs lack spatial reasoning, either in the concrete or abstract, and while that sort of reasoning has been downplayed by academia for at least a century it is fundamental to technology and industry. (And many would say for science and mathematics too).

They will, however, get there as well either directly or as interfaces to models that do, and your core point stands.

Yeah when things like this have happened on fridays before they've reset everyone's limits etc. Would not be surprising if this is the pretext for something of this sort.

If not AGI, at least greatly disturb their old cushy SaaS jobs.

I mean, this is happening right now.

I doubt there are few people talking to customers and potential customers that haven't run into the "look what I rolled last week with Claude" only for it to be legitimately impressive. The idea of technical moats is evaporating in front of our eyes.

Neither of those are beacons of either efficiency or performing to standard. In fact both are chronically mismanaged.

You'd actually struggle to pick worse examples.

Tangent: disturbing to see the BBC thinking they're going to get far by paywalling quasi-randomly in North America. At least right now it's trivial to ignore.

False dichotomies, such as your original reply, hardly count as good faith.

And now that original has even been flagged . . . this place really has tanked.

The reality is people don't always care if a human poured their heart and soul into something. Sometimes they do, but not always.

Generally speaking the ones that do care are those that also hope their own creations are/will be appreciated by people that similarly pour their heart into them, and they really don't understand that most people just see things for what they as consumers get out of them.

On some level writing on the net now is for an AI audience anyway. (Greetings fellow bots).

Casey is right, but the whole point for the Rivian position is they see themselves as the car software platform of the future. The idea we want primarily phone UIs in our cars is antithetical to their very existence, or why VW invested in them.

The idea customers should have control over the interfaces they use needs to be reinforced. If anything this is why the AI wave has gained so much hype as now people get to bypass so much nonsense masquerading as UX.

Because it converts all normal DBA tasks into development tasks. If this fits your org (and at huge scale it is more likely) then that's fine, but for small/mid/growth places that can be a PITA.

There does seem to be more effort going into an SQL layer, but it's very hard to know what the state of layer dev and direction is these days, especially if you're not located near core devs.

The one that amazes me is Okta.

OK their Mac UX is great, but given their rate of incidents how can you trust it?

Clearly this stuff is not actually bought based on track record.