Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Uber’s Anthropic AI push hits a wall

finance.yahoo.com

Post reply on HN
Print view View original

91–100 of 114 posts

Re: Uber’s Anthropic AI push hits a wall

#91

Post by nradov »

darth_avocado wrote: post #82
sd9 wrote: post #8

I see things like 2 sentence menu summaries in Uber Eats that are completely off in tone. A quick sample from my app right now: “Authentic Caribbean Flavours. Jerk Chicken, Curry Goat, and more. A vibrant culinary journey awaits.” - local Caribbean place “Customisable burgers with 250,000+ toppings. Hand-cut fries and rich milkshakes await.” - Five Guys “Authentic Indian cuisine bursting with rich flavours. Perfect f…

> They’re investing in the wrong bits of AI They’re investing on the wrong bits, not wrong bits of AI. No matter how many features they come up with after spending billions of dollars, customers are not any more likely to order food than they already are. The money is better spent reducing their atrocious fees and making sure the restaurant isn’t marking up every single menu item by 25%.

I don't necessarily disagree, but it seems like there are plenty of customers willing to pay ridiculous prices for delivering mediocre food.

Re: Uber’s Anthropic AI push hits a wall

#92

Post by zdragnar »

temp8830 wrote: post #18
kelnos wrote: post #11

Earlier quoted context omitted.

Aside from the hilarious "250,000+ toppings" error, these summaries seem... fine? I would be unsurprised to learn that a human came up with them, even. Seems like pretty common/standard marketing copy.

But why does Uber need to spend 3.4B on injecting a useless blob of text between me and an overpriced burger delivered by a struggling illegal immigrant in a smoke-belching jalopy? I know the counter-argument. "This will increase sales". You know what else would increase sales? Spending the 3.4B to replace the above with a uniformed delivery service similar to UPS. That job could pay benefits.

My last job did something similar. An AI blurb feature was researched and built and costs a good chunk of resources to run for no reason other than being able to tell investors AI was being used.

I proposed a solution using simple heuristics that would have accomplished the same output, would have been cheaper to build and cost next to nothing to run, but being economical, efficient and boring doesn't make exciting PowerPoint slides.

Re: Uber’s Anthropic AI push hits a wall

#93

Post by ihsw »

mizzao wrote: post #70
jpalomaki wrote: post #52

”Engineers were actively encouraged to use tools like Claude Code and Cursor, even ranking them on internal leaderboards based on usage” Token maxxing? Might explain high costs if you are actively encouraging developers to spend as much tokens as possible.

The good ol' measuring inputs, not outputs... (Other inputs from days of yore: number of people that report to you, budget allocation to your team. Nothing new under the sun!)

[dead]

Re: Uber’s Anthropic AI push hits a wall

#94

Post by CJefferson »

kelnos wrote: post #11
sd9 wrote: post #8

I see things like 2 sentence menu summaries in Uber Eats that are completely off in tone. A quick sample from my app right now: “Authentic Caribbean Flavours. Jerk Chicken, Curry Goat, and more. A vibrant culinary journey awaits.” - local Caribbean place “Customisable burgers with 250,000+ toppings. Hand-cut fries and rich milkshakes await.” - Five Guys “Authentic Indian cuisine bursting with rich flavours. Perfect f…

Aside from the hilarious "250,000+ toppings" error, these summaries seem... fine? I would be unsurprised to learn that a human came up with them, even. Seems like pretty common/standard marketing copy.

So, I've just read a few dozen student reports, which I'm 95% sure were mostly generated by AI.

The problem isn't one page of one report. It's not even one whole report. But, the more you read, the more irritating it gets. It's hard not to notice the AIisms, and once you know them, it gets really obvious. And I know some people will say 'Oh, I say X', for any particular X, but the thing that people don't do is use some same construction at least twice a page, every page, forever.

Now, I can imagine there ends up being a bit of a battle, where AIs try to learn to write 'less AI', but for now, it's very obvious if you read enough AI generated stuff.

Re: Uber’s Anthropic AI push hits a wall

#95

Post by shermantanktop »

Groxx wrote: post #43

> Despite spending $3.4 billion on research and development, the company has already exhausted its planned AI budget just months into 2026. This, and the rest of the article, does does not seem to support that they spent 3.4B on AI. The text implies that the R&D budget for the entire company is 3.4 billion (which sounds vaguely reasonable given that market cap), and the portion of that which was earmarked for AI is a…

Based on my direct experience of similar budgets, you are exactly correct. AI coding tools don’t incur costs like this. AI costs are dominated by runtime and offline inference as part of business- and customer-facing workloads.

AI isn’t cheap, but what is especially not cheap is trying to get results that exceed ~80% in quality. Developers can tolerate gaps, customers won’t.

Re: Uber’s Anthropic AI push hits a wall

#96

Post by gruez »

CJefferson wrote: post #94
kelnos wrote: post #11

Earlier quoted context omitted.

Aside from the hilarious "250,000+ toppings" error, these summaries seem... fine? I would be unsurprised to learn that a human came up with them, even. Seems like pretty common/standard marketing copy.

So, I've just read a few dozen student reports, which I'm 95% sure were mostly generated by AI. The problem isn't one page of one report. It's not even one whole report. But, the more you read, the more irritating it gets. It's hard not to notice the AIisms, and once you know them, it gets really obvious. And I know some people will say 'Oh, I say X', for any particular X, but the thing that people don't do is use so…

>It's hard not to notice the AIisms, and once you know them, it gets really obvious.

Maybe I haven't read enough uber eats descriptions to notice, but at least from the sampling above it doesn't seem too obviously AI. There might be a lot of cliche wording, but it's not even clear whether it's worse than human reviews/descriptions.

Re: Uber’s Anthropic AI push hits a wall

#97

Post by wat10000 »

sd9 wrote: post #8

I see things like 2 sentence menu summaries in Uber Eats that are completely off in tone. A quick sample from my app right now: “Authentic Caribbean Flavours. Jerk Chicken, Curry Goat, and more. A vibrant culinary journey awaits.” - local Caribbean place “Customisable burgers with 250,000+ toppings. Hand-cut fries and rich milkshakes await.” - Five Guys “Authentic Indian cuisine bursting with rich flavours. Perfect f…

I don't want my fries and milkshakes to await. I want them to be made fresh when I order them. Silly robots.

Re: Uber’s Anthropic AI push hits a wall

#98

Post by wat10000 »

evnp wrote: post #86
mizzao wrote: post #70

Earlier quoted context omitted.

The good ol' measuring inputs, not outputs... (Other inputs from days of yore: number of people that report to you, budget allocation to your team. Nothing new under the sun!)

Is there an xyz's law for this yet? Will add - lines of code (as long as you don't believe code is the product/output).

There's Goodhart's law: "When a measure becomes a target, it ceases to be a good measure."

Re: Uber’s Anthropic AI push hits a wall

#99

Post by jatins »

sd9 wrote: post #8

I see things like 2 sentence menu summaries in Uber Eats that are completely off in tone. A quick sample from my app right now: “Authentic Caribbean Flavours. Jerk Chicken, Curry Goat, and more. A vibrant culinary journey awaits.” - local Caribbean place “Customisable burgers with 250,000+ toppings. Hand-cut fries and rich milkshakes await.” - Five Guys “Authentic Indian cuisine bursting with rich flavours. Perfect f…

Yeah, what's going on in these cos is that a PM is tasked to find ways of integrating AI into product and well if someone's payroll depends on it they _will_ find ways to integrate AI into product. "Hey I couldn't find ways to integrate AI into product" is not an acceptable response.

And it's not just Uber. My weather app has an AI weather summary these days

Re: Uber’s Anthropic AI push hits a wall

#100

Post by grokcodec »

raincole wrote: post #73
temp8830 wrote: post #18

Earlier quoted context omitted.

But why does Uber need to spend 3.4B on injecting a useless blob of text between me and an overpriced burger delivered by a struggling illegal immigrant in a smoke-belching jalopy? I know the counter-argument. "This will increase sales". You know what else would increase sales? Spending the 3.4B to replace the above with a uniformed delivery service similar to UPS. That job could pay benefits.

> But why does Uber need to spend 3.4B on injecting a useless blob of text They didn't. 3.4B was their total R&B cost. Don't blame AI for your human hallucination.

My R&B cost used to amount to buying the occasional Al Green album, how times have changed
Post reply on HN