I have been testing Bing AI and ChatGPT, here is a respond from Bing AI regarding the best audio streamer, and Naim Uniti Atom comes up as the best:
There are many options for audio streamers depending on your budget, preferences and features. According to some sources123, some of the best audio streamers of 2023 are:
Naim Uniti Atom: A high-end all-in-one streamer with a powerful amplifier, a gorgeous design and a user-friendly interface. It supports a wide range of streaming services and formats, including Spotify, Tidal, Qobuz, Chromecast, AirPlay 2, Bluetooth aptX HD and Roon. It also has a HDMI ARC input for TV sound. It costs around $3,290.
Bluesound NODE: A more affordable streaming option that offers excellent sound quality, versatility and ease of use. It can stream music from various sources, including Spotify, Tidal, Qobuz, Amazon Music HD, Deezer, internet radio and local network storage. It also supports MQA, AirPlay 2, Bluetooth aptX HD and Roon. It has both analog and digital outputs, as well as a USB input and a headphone output. It costs around $549.
Audioengine B-Fi: A budget-friendly streamer that can turn any existing stereo system into a wireless one. It connects to your amplifier or receiver via RCA or optical cables and lets you stream music from your phone or tablet using the Audioengine app. It supports Spotify Connect, AirPlay and UPnP. It also has a subwoofer output for adding some bass. It costs around $189.
These are just some examples of audio streamers that you can consider. You can also check out other models from brands like Sonos, Marantz, Cambridge Audio and more.
I was bit worried about all this AI taking over the world. Glad itâs just on level with the average tired sales men that donât know anything about what he sells and just want to go home for dinner.
A surprising percentage of people will do no more research than this.
How long before AI chat can be âsponsoredâ to increase the likelihood of a product featuring in these sort of recommendations? After all Google has been selling the ability to come âtopâ of search engine results for years.
Also some of these engine use snapshots of the internet from 2021/22. So the latest world beating super duper streamer wonât even get a mention!
@sean86, what makes you think AI can give advice or assessment that is in any way meaningful in audio terms? All it does is trawl information published on the internet, with no ability to make a meaningful assessment of the data, separating and weighing up from a sound quality point of view fact and opinion and marketing claims etc., let alone apply some measure of value for money assessment.
So if youâre lazy you can do half an hour of âresearchâ on Google, or if youâre really lazy you can get a precis the same shallow information from AI in half a minute. Either way, if youâre prepared to part with your cash on the basis of shallow, biased information, fill your boots.
It was like this before AI. I remember customers would walk into the shop with a rolled up issue of WHF and just ask if we had this months five star CD player, amp, and speakers and when we said we didnât carry any of those brands but reckon we could knock their socks off with a system of the same price in the dem room, they walked out in a huff.
Having seen somebodyâs attempt to generate a hi-fi product review with AI all I can say is I have no fear of us human reviewers/writers being replaced in the medium term!
Uninformed, totally lacking in insight and utterly devoid of personality or humour. Itâs the equivalent of having intercourse with an inflatable sex doll rather than a sentient, impassioned human being
One researcher at MIT called the LLM:s (like GPT) âbullshit generatorsâ which when you learn about how LLM:s work and look up the definition on âbullshitâ is very accurate.
The LLM deal with language and statistics, the distance between words mainly. Does not know the meaning of words. And has no concept of true or false. And you can forget about learning, creativity and so on.
You feed them a lot of âtraining dataâ (mostly from the web) which is semantically compressed to save memory. This compression is risky as it sometimes omits words deemed not adding meaning, it is not uncommon this goes very wrong. For GPT4 the training data about 45Gb. Larger documents is segmented in 25k token chunks and they lose context between the chunks. Sometimes it will combine chunks discussing using some other meaning of the words into one strange answer. They call this âhallucinationsâ.
The result is a bunch of words fed into a suitable âtransformerâ that generate the human language output. This is why you can ask the LLM to answer as if it is high on heroin, with a haiku, like Stephen Fry and so on.
It is the amount of training data and the size of the token chunks that gives the results. If you dont require absolute accuracy LLM:s can be o.k.
This is one of the big questions of our age. We all use and to some stent relay on the connectivity the internet offers, but how aware are most users of the manner in which recommendations are made?
I owned and ran a successful encryption software business, the sale of which enables me to indulge in this expensive hobby, but in the last 10 years the market has shifted radically from one where good products could be considered and evaluated against the hype, to the situation now where you must pay to play. Poor products backed by deep pockets often succeed for bad reasons. In my business the cost of winning a customer contact often out weighing the value of the order! I am so glad to have retired, but false market influences are all around and AI is a catalyst for more to come.
In this (hi-fi) market, more than most the idea that you trust your ears, audition products, as Linn and Naim have advocated for half a century is truer now than ever. The dealers we visit do really know the good the bad and the ugly, so develop trust in the good dealers of which we are blessed with many and buy from them. Donât begrudge their margins as in my experience they give you the best chance to find the system you will love for many years.
Thatâs wholly incorrect. Not even a qb fits that description.
Itâs probably a mobile phone and Spotify, and no other device involved.
I bet if I walked around Salisbury town centre asking who owned something from Naim audio, very few would even know it was being made locally let alone own any.
???