Hacker News Reader

Comment by reader9274 | Hacker News Reader

I think we're all reaching AI fatigue. Fewer and fewer people care anymore

Sure if you're viewing this as some kind of spectator thing, or entertainment, maybe it's less interesting. But it doesn't really matter whether "people care". What matters is whether it's useful and has impact. It's enough if the small number of people use it for whom it is useful. It doesn't matter if the average Joe on the street is excited by it.

Few people care or even know about various advances in various specialized fields. It's enough if AI simply seeps into various applications in boring and non-flashy ways for it to have significant effects that will affect a wider range of people, whether they get hyped by the news announcements or not. Jobs etc.

An analogy: the Internet as such is not very exciting nowadays, certainly not in the way it was exciting in the 90s with all the news segments about surfing the information superhighway or whatever. There was a lot of buzz around the web, but then it got normalized. It didn't disappear, it just got taken for granted. No average person got excited around HTML5 or IPv6. It just chugs along in the background. AI will similarly simply build into the fabric of how things get done. Sometimes visibly to the average person, sometimes just behind the scenes.

Not sure if it's just me, but it looks like all SOTA companies are doubling down to chase the new benchmark, which beyond hype, doesn't seem to translate into real world uses. Why don't these companies just plug it into a popular git repo and say, hey our AI fixed these 100 issues! Or something real? The only people who seem to be doing something real is DeepMind.

Incorrect. We are not all reaching AI fatigue.

Especially this is not a breakthrough justifying a 340B USD valuation, but rather the work that junior developers can do; implement a loop of Bing Searches connected to an LLM.

Peak HN comment

Doesn't make it untrue.

Agents that can search the internet exist for a while now and have been essentially solved and happily used in platforms like Perplexity.

It's really "meh", very far from revolutionary.

Keep in mind this company is trying to convince everybody they need 500B USD now (through the Stargate project).

To go from partially automated to fully automated is thousands of non trivial edge cases and unforeseen decision points that must be tamed.

To say this is trivial is like saying the one shot ai prompted twitter clone is the same thing as twitter.

Peak HN indeed.

Let us know when your Bing-bot scores over 20% on the HLE benchmark.

It's literally a browsing agent that searches the internet and they know the questions in advance when preparing the agent

Without internet: 10%

With internet: 23%

In addition:

> We found that the ground-truth answers for one dataset were widely leaked online

in very small letters, and they blocked these URLs at runtime but not training time.

It's not bad, but not revolutionary at all compared to the leap that was GPT-2 from GPT-3, or GPT-4o to DeepSeek-R1

If they "knew the questions in advance," why'd they need Internet access at all? The ability to use the same data sources humans would use is not the insult you seem to think it is.

Again: the assertion was yours, so let us know the results of your own work.

I haven’t tried the OpenAI version yet, as I’m on their peasant-level $20 plan, but the Google equivalent is way superior to Perplexity (I use both extensively). The web search Perplexity carries out is superficial compared to the Google product; it misses a large percentage of what Gemini Deep Research finds, and for a particular task in my business this makes a huge difference.

Reply on news.ycombinator.com