Live data from Hacker News

Claude can now search the web

anthropic.com

401–410 of 758 posts

Re: Claude can now search the web

#401

Earlier quoted context omitted.

Grok's DeepSearch and DeeperSearch are also pretty good, and you can look at their stream of thought to see how it reaches its results. Not sure how OpenAIs version works, but grok's approach is to do multiple rounds of searches, each round more specific and informed by previous results

My disgust and hatred for Elon Musk prevents me from giving Grok a fair chance. I realize this is my psychological problem. I own it, but as far as I can tell, I'm not missing much.

It is paramount to a functioning society to have zero tolerance for nazis.

Re: Claude can now search the web

#402
post #365

Earlier quoted context omitted.

We don't just have "bullshit jobs" (which is an actual term these days), we have a "bullshit economy" as well - centered around advertising because without advertising most of the bullshit just wouldn't sell. Like, if you already got a car, you can drive it for 10-20 years easily, or more if you take well care of it. But advertising makes you think you "need" a new car every few years... because that keeps the econom…

Cars from 20 years ago emit significantly more polluting substances. OTOH they are lighter weight and thus wear the roads less. On the third hand, none of them is electric or hybrid. Some computers from 20 years ago are still in a good shap, but... (You can continue.)

I think this is a different argument than the disposable, single-use economy being described.

The volume of things we buy but don't need (or necessarily want) drives a huge sector of the global economy. We're working to fill our lives with unnecessary things that bring us no happiness beyond the adrenaline hit when we hit "Buy Now" and the second one when the Prime box arrives at our door.

Consumerism masks the underlying problem and it's only going to get worse as more is automated. Producers will have an incentive to convince us we still need more.

Cars are - to me - a red herring in this argument except for the people who do literally trade in for a new car every few years. I drive whatever fairly boring Honda for as long as I can (usually 8-10 years) and don't feel a ton of regret about investing in comfort. But I've been as guilty as anyone about just buying stuff because it pops up in an ad or recommended on Amazon, etc.

Re: Claude can now search the web

#403
post #368

Earlier quoted context omitted.

To be fair, being in the cultural zeitgeist is a huge part of their current moat. To people in the street OpenAI is the company making LLMs. Sam has to make sure it stays that way

Fake it til you make it vibes. I understand why he would do it, what I find strange is that it inspires confidence in the market.

You can wait to release something all over the world which will take time because it’s not an engineering issue but a compliance/legal or other types of issue. Or you can iterate faster by doing the minimum and getting feedback and then releasing it in other markets. Not sure what’s wrong with this approach.

Re: Claude can now search the web

#404

Earlier quoted context omitted.

Compared to OpenAI, who seem keen to maintain the mindshare of everyone, IMO Anthropic are far more considered about their audience. They released a report recently on who who was using AI professionally and it was something like 40% developers, and single digit percentage for basically every other profession. I think they’re focusing on the professional use cases.

Pretty much. Claude from their announcements seems to me at least to be about SWE's and coding at the moment. Personally while I understand their decision I find it a bit limiting, and just a little targeted against the SWE profession. If all AI does is disrupt SWE's but not really add new products and/or new possibilities; then it feels IMO like a bit of a waste and is quite uneven in its society disruption. At leas…

I know people in other industries use AI a lot and likes it. Accounting, legal, writing (a lot here). I agree that companies that focus on all verticals like openai is definitely the way to go. Claude code capabilities are not very significant compared to openai though. There is no big moat and a lot of it is perception, marketing.

Re: Claude can now search the web

#405
post #382
post #343

Earlier quoted context omitted.

One suggested weakness of UBI is a lack of purpose. I wonder if the "solution" is somewhat as you implied: jobs without a strict return on investment. You get your stipend, but you're keeping your block clean by sweeping and mulching. They're getting theirs in exchange for cranking out sourdough at cost for the neighbourhood. Someone else gardens for elderly residents.

Not UBI per se, but this exists in rural parts of Southern Spain in some way, and is called Rural Employment Plan (PER in its Spanish initials). The give simple jobs, like cleaning or painting, to people on the lower bottom of earnings. Most people in that plan are people with low formation, like those who left school in their mid teens.

Sounds like a minimum wage

Re: Claude can now search the web

#406

Does not really say /how/ it's performing a web search... Is it tapping into it's "own" corpus of material or calling out to some other web search engine?

In my quick experiment (asking a question that would naturally lead to content on my own site) it is not doing a real time request to the site in question. Its answer included links back to my site (and relevant summaries), but there was no requests for those pages while it was generating its answer. So it's clearly drawing from info that has already been scraped at some earlier point. And given that I see Claudebot…

Major AI players don’t want to use someone else web index as they may cut it off or jack up the prices etc. major players want to build their own web index

Re: Claude can now search the web

#407
So in many respects, search the place it used to construct the model? Isn't that functionally bias-reinforcing?

"Look what I synthesise is correct and true because when I use the same top 10 priming responses which informed my decision I find these INDEPENDENT RESULTS which confirm what I modelled" type reasoning.

None of us have a problem with an LLM which returns 2+2 = 4 and shows you 10 sites which confirm. What worries me is when the LLM returns 2+2 = 5 and shows 10 sites which confirm. The set of negative worth content sites is semi infinite and the set of useful confirmed fact (expensive) sites is small so this feels like an outcome which is highly predictable (please don't beat me up for my arithmetic)

e.g. "Yes Climate science is bunk" "

Re: Claude can now search the web

#408
post #383

Surprised that Claude (the app, not model) not only has done well for so long, but has somewhat consistently clinched the top spot in coding, all without a feature that is considered somewhat of a basic feature for most consumer-facing AI apps.

How much % it’s significant compared to say openai or google? Because if I’m paying 20$ I want other things too not just coding. And if the moat for coding compared to other vendors is not significant, it doesn’t make any difference tbh

Re: Claude can now search the web

#409

i stopped using Claude about 2 months ago. went to Grok (the code was better, everything was better - politics aside). i wonder if this update will improve it. the main issue i find with Claude is, he fights you. He refuses so many requests and i need 3 or 4 replies to get what i want vs deepseek/grok. i've kept the monthly subscription to help anthropic, but it's trounced by the free options imo.

Claude 3.7 with thinking is a big step up. If you haven't used it I'd suggest giving it a try. I have used grok a bit and it did what I needed it too, so I can't really compare. But 3.7 thinking is crazy strong for coding.

I got much better results with o3-mini than Claude for coding. Get the high level design wkth o3 or o1 and then use gpt4 or Claude for small tasks.

Re: Claude can now search the web

#410

Earlier quoted context omitted.

UBI will almost certainly fail to cover the necessities unless we have Marx-style price controls. When everyone in the economy has a minimum of say $3,000 per month the cost of necessities, and everything else, will go up roughly in line with that.

Case study: COVID

That isn't a test of anything, since we've not isolated a single policy change; many things changed everywhere all over the world all at once.
Post reply on HN