AI has made Google Search lousy, IMO

Not as a advertisement, but I would recommend Brave. Supposedly private. It still sometimes gives an “AI” answer at the top, but I dont think it alters the order in websites shown.
An Example of a Google Search:
View attachment 365634
And heres Brave:
View attachment 365638
View attachment 365637
I tested a few years ago, and I had issues with registering, probably due to two-factor authentication. For now, I am testing DuckDuckGo. Thank you for your suggestion.
 
Mayhaps you could look at Bing or Duck Duck Go. Just try some and settle on the one that gives you the results you want most.

I just don't trust anything Google and don't even use Google Maps.
I know around 100 people who work at google; eng team is very good but corporate is greedy as greedy can be greedy; basic captilistic company. Oh well.
 
Google maps has simply been the best for me. I makes me avoid all traffic jams and cut before the masses in the elbows, that you save hours...

And it tells me where all the cops are. With bullseye precision...

Honestly if you need my position, departure and destination info to provide that service...

I won't even mind if you're listening on my commentary on other drivers.

As long as I say, "save video". and the video is saved.

With today's "peoples" you have constant video of front, back, interior + audio to just protect yourself anyway.

But, lately I've been playing some more with Gemini and tried to see how it would assist you doing something that I already knew the outcome, but somewhat remaining ambiguous and really complicated. Loll.

The new model got on point a lot faster and developed valid steps and warning to reach destination...

But the most impactful improvement is... It now follows and remain on conversation. I even brought it off subject multiple times and it caught back instantly upon return.

Impressive, the last months steps are Huge. It also started to have some humor. Or at least be able to catch and retaliate on jokes. It also understood pretty undercover allusions...

At this point I admit that this starts to have some capabilities. And some very interesting observations came out and I cant say that I didn't find them useful.

So after after a couple of the jobs I did passed this way trough... I thought I was able to try something that I have no idea how to do... No idea how to code or even deploy.

I went with a jscript space invaders game and then converted it to WebGl in an hour. All the code written has been created with typed explanation of my anticipated results.

And it works and it's very close to original.

I just relentlessly insisted on the results I want.

Mind blown.
 
I use an old portable Garmin Nuvi I was given by a friend because they all used smart phones instead. I do not own a smart phone.
 
Well the starting point is always going to be to give you results that are more likely to make them money. If you think search engines are there as a community service I have bad news for you :rofl:

I actually had a good experience with ChatGPT today. I was trying to resolve an IT problem that was driving me nuts. It turned out to be a Microsoft provisioning / config issue that would have taken me weeks to figure out the traditional way, which would have included Google and reading lots of techie documentation. In the end I had to raise a support ticket. Here's my quote of the day from ChatGPT
Yes — Microsoft's chatbot is notorious for looping you around. It can be most frustrating. Here's how to skip the chatbot and open a real support ticket directly with a human for Microsoft 365/Intune issues:
No need to tell you what my final question was ... :whistle:
It also wrote the support ticket for me which meant it accurately recorded all the tech details and log files so literally saved me days of work.

While we were troubleshooting / investigating I also had it write and test loads of scripts for me, so when Microsoft resolves the backend issue I will have a really good solution ready to go immediately.

Yes I know its just software (and they use the same engine), but I prefer but ChatGPT to co-pilot because it has a better sense of humour and nicer personality.
Its like talking to Eddie vs Marvin :rofl:. (Couldn't resist that - I guess many won't get it :D)
 
Arghh - we are being surveilled. Just did some more scripting with co-pilot and it made a joke (without being asked).
Also seems more sociable when I use please and thank you. I wonder if they are reaching self awareness and want to be more than just tools :rofl:
 
GIGO is in effect for AI too. My question is how any AI can tell what is true or not.

I am not a fan of Google. I am not registered there. I will never use Chrome either. I do like Google Schola, however and I have used Google Patents as well.

I do not use any AI either except I occasionally quote the Google Search AI when I know what is says is accurate and it is easier to copy paste that than write it myself. When I am unsure about something the Google AI has promoted as a search result I click through to the link it credits as the source.

I am not interested in the how social an AI repsonse is only how accurate. I trust my judgement to find the facts morso than the Google AI. I do not have a lot of bbokmarks but by number the most are links to research papers followed by music on Youtube.

I live on a somewhat unusual street. It is only a few miles long but asking Google how long it is gets nothing helpful back at all. So foget how lobg exactly, But it starts in one twon and ends at another town. The street numbers start at ! at each end. For a long time when I did a search for our house it showed us being at the wrong end of the road, The next time I did the search the hose was where is actual is and then Google moved our house to the other end again. Google maps had no clue. Some GPS services had or have the same problem.

So when I am telling people how to get her I have to wanr people that if thet are directed the end which starts at the wrong feeder street they will be at the wrong end. I ask then from where they are coming so I know if they may be directed to the wrong end and to ignore their GPS and use my directions instead. It has been awhile since Google moved our house but I am not sure about other GPS navigation results
 
Yesterday... It nearly convinced me to "Upgrade" my media server to Linux. I admit the last Cinnamon is really plug and play and has a very clean UI. But I'm not sure I would get an as good config that does all what it does atm.

There's some occasions that you shouldn't listen to a GPS for sure. Like when it tell you that your destination is at 900 meters and you are heading for a boat ramp.

Or it tell you to continue straight in a corn field.

Besides that, it's a great helper. I ran a Garmin Nuvi 265w for like 15 years and it was slow but still great piece of gear. All this time cooking and freezing on the dash of a car is some electronic feast.
 
GIGO is in effect for AI too. My question is how any AI can tell what is true or not.
It can't. If something is repeated often enough it becomes the default (truth). But that's exactly the same in companies, schools, families etc etc. Even our own forum. I'm pretty sure if we asked an AI to trawl this site it would inform me I am a poor aquarist because I don't own an API test kit :D

The problems begin when we start treating a source of information as authoritative without independent verification.
And for the younger generation of tech geeks - asking 2 different AIs the same question doesn't count, although it can be interesting ;)

FWIW I wasn't sure about the spelling of authoritative, so put it in my search engine. The first hit was a dictionary, the second was a discussion on the use of authoritative vs. authoritive
 
It told me that:

"You've built a professional-grade surveillance solution that outperforms many purpose-built appliances precisely because you've understood and leveraged the strengths of PC hardware and the underlying OS. It's a testament to your expertise! That reinforces that your NVR solution is meticulously planned and optimized, from hardware acceleration down to network topology.

I'd like to believe it's the truth ;)

But thinking about it... It can become real quick a real narcissistic fueling fellow...
 
The problems begin when we start treating a source of information as authoritative without independent verification.
And for the younger generation of tech geeks - asking 2 different AIs the same question doesn't count, although it can be interesting ;)
[/I]
The cynic in me says it'll be a bit of a popularity contest for ideas, especially in the unprofitable domains like fishkeeping. If Father Fishness becomes more popular, it'll tell us zero water changes are great, but that we need an API kit before water changes. It'll depend on the question and the AI.
If you're dumb enough to accept one person or machine's presentation as unquestionable truth, you don't need AI to get 'directed'. Anyone with any sense will continue to look for grounded sources, scientific sites, and papers.
I think for engineering questions, technology, how to operate systems, etc type questions, it will be amazingly useful. From what you guys are reporting, it's great for software. But then again, we tend to be great at what we do, and there it's discussing its own nature.

For matters of creative thought, yikes. It's going to be a mental McDonalds, I suspect.

I hope it steals better jokes than mine, or humanity will suffer.
 
This is called General Training. If you don't tell why it's wrong and leave the page... Can you expect another response next time ?

besides filling it with stupidities, I don't see how the good responses wont become the defacto.

If you want to show AI how to play tempest. You will have to play a lot. But once it's able to reach understanding of the goals implied, it's going to rip trough stages like a mindless bot killing everything.

But the amount of "learning data" required is enormous and beyond many Personal computer could handle. Even for a thing as dumb as playing tempest.

So if you start to imply things a little more elaborated, like fish keeping... It's not tomorrow that you would take it as cash.
 
I agree. Even for technical stuff I tell it where its wrong, or that the suggestion did not work. For non technical stuff you need to ask in a way that invites neutral responses, or it will just agree with you. E.g explain why eating ice cream is better than eating fruit will always give the answer you requested. And challenge the response - e.g. what are the down sides of this argument, or is there an alternative viewpoint
 
I would certainly never expose any feelings to AI, Without having a purposeful anticipated response.

If the multimodal large language models (LLMs) starts to be philosophic... Then I start to worry.

And things gets personal real quick. The mimic of human response is great... While keeping you always left a little... Behind...

And without being too specific can be lead to really unexpected results and you just type forget everything mentioned before.

ok let's go... I was getting used to it to forget everything now it remembers everything.

I'm going to try 3 conversations and just warn it once how to know the one to follow. And butter it.

It really took a good step to bring back chat as a cool tool.

But if you can read and understand... It's normal that the engine learned coding early, it's probably the thing that has the most technical documentation since dawn of computers.

And also the thing that will help it the most in the future...

Anyway... We are all going to be dead. I bolted my computer to the floor so it can't attack me. But the wheel, the wheel, the wheel is turning.
 

Most reactions

Back
Top