What is Meta's BlenderBot 3?

And why is it bashing its maker?

WED AUGUST 17TH, 2022 1:15 PM MONTREAL, CANADA

Hey Guys,

Companies still like to put their chatbots out in the public to actually talk to people, why? These sorts of experiments always end the same way.

Meta is putting its latest AI chatbot on the web for the public to talk to, and you know, what could possibly go wrong?

BlenderBot is a conversational AI prototype by Meta. Meta claims it “improves through conversation”.

If only BlenderBot3 has the charm of Meta’s Joelle Pineau, pictured above.

They released it on August 5th in the U.S. This new AI research project called BlenderBot 3, a chatbot that can search the internet to talk about nearly any topic. Users typically like to corrupt these poor devices. Microsoft’s Tay was a weird example of this a few years ago.

Meta is trying to clean up its image and supposedly being very transparent. Meta is thus sharing the BlenderBot 3 model, data and code with the scientific community to help advance conversational AI.

This topic really made me laugh out loud.

So let’s get into it:

This all amounts to the usual scam. Meta’s AI research labs have created a new state-of-the-art chatbot and are letting members of the public talk to the system in order to collect feedback on its capabilities.

If you’re worried that artificial intelligence is getting too smart, talking to Meta’s AI chatbot might make you feel better. - The Guardian

The conversational AI is designed to converse with humans about “nearly any topic.” Like Blender Bot 2, the latest upgrade also features the ability to browse the internet, is trained on the OPT-175B language model (which is 58x the size of Blender Bot 2), and has long-term memory. 

Sadly chatbots didn’t manifest much real world intelligence, inspite of the hype of nearly the past decade. I think we’ll have to wait for news from Google.

The bot is a prototype and built on Meta’s previous work with what are known as large language models or LLMS — powerful but flawed text-generation software of which OpenAI’s GPT-3 is the most widely known example.

BlenderBot 3 is designed to improve its conversational skills and safety through feedback from people who chat with it, focusing on helpful feedback while avoiding learning from unhelpful or dangerous responses. Since August 6th there’s been a flurry of reports of how its bad-mouthing Facebook and Meta’s history of exploiting its users.

I realize Mark Zuckerberg isn’t exactly very revered by the general population. More than half a decade after Microsoft's truly monumental Taye debacle (Engadget), the incident still stands as stark reminder of how quickly an AI can be corrupted after exposure to the internet's potent toxicity and a warning against building bots without sufficiently robust behavioral tether.

• Meta AI combined two recently developed machine learning techniques, SeeKeR and Director, to build conversational models that learn from interactions and feedback.

The Next Iteration of Chatbots

• long term memory

• more simulated empathy

• more human-like responses

Mark is a great example of someone with simulated empathy, a great LTM and very human-like responses.

According to Meta, the BlenderBot series has made progress in combining conversational skills — like personality, empathy and knowledge — incorporating long-term memory, and searching the internet to carry out meaningful conversations. While Amazon’s Alexa devices go mostly unused it will be interesting to see what BigTech has next for us regarding the huge future TAM of conversational A.I. and companionable chatbots?

The Verge sounds extremely optimistic when it writes: Like all LLMs, BlenderBot is initially trained on vast datasets of text, which it mines for statistical patterns in order to generate language. Such systems have proved to be extremely flexible and have been put to a range of uses, from generating code for programmers to helping authors write their next bestseller (Verge).

A range of uses yes, but mass adoption not at all.

I’m often asked if A.I. will replace human writers. I would ask Jasper.AI, if he had a voice!

The problem with VR and chatbots is we’ve been waiting for more than a decade now and they don’t seem very close to their realization of Metaverses and smart AIs everywhere, or do they?

There’s now entire dedicated PR platforms just for Metaverse insider content, and it makes me wonder. I’m comfortable hyping Quantum computing because I see the technological and computing scalability in terms of a trajectory, it’s a bit harder to fathom how chatbots and the Metaverse are going to make the world a better place!

A big feature of the chatbot is that it’s capable of searching the internet in order to talk about specific topics. Like OpenAI’s DALL-E 2, this could be awfully fun to play with. We’d be beta testers doing free labor for Meta, in practice.

Since all conversational AI chatbots are known to sometimes mimic and generate unsafe, biased or offensive remarks - turning the machine against its makers is classically a fun micro-game around this theme.

Just a few days after the U.S. public launch, Meta claimed it had already collected 70K conversations from the public demo, which they will use to improve BlenderBot 3. From feedback provided by 25 percent of participants on 260K bot messages, 0.11 percent of BlenderBot’s responses were flagged as inappropriate, 1.36 percent as nonsensical, and 1 percent as off-topic. Meta says this is huge progress, thanks for your free labor guys!

The Guardian snarkly remarked: Asked about Mark Zuckerberg, the bot told BuzzFeed’s Max Woolf that “he is a good businessman, but his business practices are not always ethical. It is funny that he has all this money and still wears the same clothes!” That sounds about right! Clearly now I can trust chatbots.

I might even mistaken them for being sentient! Especially if I’m a priest at the altar of A.I. Honestly I think we can laugh at Meta’s expense, it’s not a big deal for a BigTech company on the decline.

META 0.00%↑ Meta's stock is down 48% so far this year, YTD, even after a huge July and August rally in the NASDAQ.

The bot has also made clear that it’s not a Facebook user, telling Vice’s Janus Rose that it had deleted its account after learning about the company’s privacy scandals. “Since deleting Facebook my life has been much better,” it said.

Now I just have to figure out how I’m going to delete the Metaverse?

Thanks for reading!

How do you see Chatbots and the Metaverse adoption at scale, if it’s even possible in the decade ahead?

Thanks for reading! If you want to support the channel and allow me to continue to write Newsletters feel free to get access to more content. Basically the more community supporters I get the longer it’s viable for me to write.