• fv/
  • Posts
  • šŸ¤– GREAT EXPECTATIONS

šŸ¤– GREAT EXPECTATIONS

Thereā€™s a secret to interacting well with AI. Itā€™s the same as the secret to interacting well with people


Covering culture, tech, luxury, travel, and media in 5 minutes, 3x a week

Futureboard

The Petit Pho at Pho Tai in Paris (Pete Kiehart/Eater)

šŸ”¼ New York Cityā€™s leisure and hotel revenues. Theyā€™re on pace to exceed pre-pandemic levels Ā»Ā»

šŸ”½ In house content creators. BuzzFeed is shifting to a model that outsources the siteā€™s content creation to independent creators Ā»Ā»

šŸ’¬ ā€œPeople have a natural intuition aboutā€¦ perspective.ā€ Scientists have figured out why selfies are so popular Ā»Ā»

šŸ›« Parisā€™s 38 essential restaurants Ā»Ā»

šŸ‘— Tom Ford released his final collection, made up of reissued archival pieces, in a series of videos by Steven Klein. Is he done with fashion? Ā»Ā»

šŸ’Ž This 2020 Gulfstream G500 just had its price reduced to US$40,950,000. Itā€™s only logged 453 hours of flight time since new, and has an onboard espresso machine Ā»Ā»

Quickfates

The raw and gallery-esque interior of the first ever Aries store in Londonā€™s Soho (Louise Melchior/Aries/Wallpaper)

ā€œI work remotely from vacation rentals. It's cheaper than rent or a mortgageā€ Ā»Ā»

MSCHFā€™s dating website is a glimpse at beauty standards through the eyes of AI Ā»Ā»

Inside the lucrative new business of selling AI prompts Ā»Ā»

Hereā€™s your first look at Prince William and Kate Middleton in The Crown Season 6 Ā»Ā»

Open now in Londonā€™s Soho, cult streetwear brand Ariesā€™ first permanent home doubles as an art space, community hub and cafĆ©. Pro tip: I visited the space with a longtime FATE V FUTURE subscriber yesterday; the coffee comes courtesy of the hipster Paris cafĆ© Paperboy, and is excellent Ā»Ā»

Tech jobs arenā€™t returning Ā»Ā»

Ad agency group WPP expects AI to become a 'fundamental' element of its business Ā»Ā»

A British wildlife group is looking for someone to work on one of the remotest islands in the world for 13 months Ā»Ā»

Snap shares dropped 20% after hours Thursday as the company reported 1Q results that missed analystsā€™ revenue expectations Ā»Ā» This newsletter has previously questioned Snapā€™s whole model Ā»Ā»

Meanwhile, Amazon shares surged Ā»Ā»

@socks_house_meeting is Instagramā€™s purveyor of memes about hyper-local London trends. He never imagined his account would become a widely shared staple of niche British internet humour Ā»Ā»

Meet the woman behind the ā€œNo Context Successionā€ meme account. It has over 250K followers, including the showā€™s creator Ā»Ā»

2 ex-a16z engineering and security heavyweights are starting a crypto custody firm Ā»Ā»

A new high-speed train could take you from LA to Vegas in 2 hours Ā»Ā»

Mercedes released a nice looking new E Class Ā»Ā»

Also: Mercedes EV owners in North America who want a little more power can now buy some extra horsepower ā€”for a monthly fee Ā»Ā»

Director Greta Gerwig said her upcoming Barbie film was inspired by disco Ā»Ā»

The ancient trick to help you think more wisely Ā»Ā»

A new app called Studio AI will let you design a website by describing what you want into your computerā€™s microphone Ā»Ā»

Deepfate

Thereā€™s a secret to interacting well with AI. Itā€™s the same as the secret to interacting well with people

Snapchat saw a spike in 1-star reviews as users panned the appā€™s new ā€˜My AIā€™ feature (Snap)

THIS NEWSLETTER HAS previously said that you shouldnā€™t worry if you don't use ChatGPT yet, because generative AI is coming to you

Actually, it already has.

Snapchatā€™s ā€œMy AIā€ bot was initially celebrated as the easiest, most seamless way to interact with ChatGPT. Those good feels turned sour fast when Snap decided to pin ā€œMy AIā€ to the top of everyoneā€™s chat screen ā€”yep, even above their real friends, and their most recent communicationsā€” which annoyed many users.

Elsewhere on the Internets, Google made a big splash recently about how generative AI is coming to its professional software suite, Workspace. A slick video showed Googleā€™s version of ChatGPT automating everything from well designed sales presentation decks, to poetic team emails.

NOW META WANTS to bring AI to all of its products, and Wall Street loves it

Well, technically investors were loving the US$28.6b in revenue Meta reported, as well as its 2b daily Facebook app users ā€”a record. (Pro view: the news wasnā€™t all amazing; Metaā€™s quarterly profit dropped 24% year on year, to US$5.7b.)

But while Facebookā€™s quarterly revenue and user numbers are impressive by any measure, the much bigger deal is the fact that Facebook's founder has (finally) fully shifted his attention from the Metaverse to AI.

Mark said that generative AI would end up touching every single one of Meta's products. Let's explore how that might work with just one of ā€˜em, shall we?

SEAMLESSLY INTEGRATING AI into WhatsApp is a huge deal

The reason is simple: itā€™s really, really hard to get a lot of people to try new software. (Yes, ChatGPT was the fastest ever product to reach 100m users, but zoom out: thatā€™s still just 5% of Facebookā€™s 2b users.)

Baking generative AI into an app people already have, like, and use ā€”like WhatsAppā€” should exponentially boost product adoption.

ALSO: THE WORLDā€™S businesses are about to have an infinite number of customer service reps

As Zuck put it, ā€œI expect that a lot of interest in AI agents for business messaging and customer support will come ā€”once we nail that experience.ā€

This is also (potentially) huge. Imagine never having to listen to hold music ever again, because you can simply text your airline, your bank, or the IRS to chat through a complicated issue in real time. It sounds too good to be true. But that's exactly what baking generative AI directly into WhatsApp should bring.

IS THERE A downside?

Maybe. Seamlessly integrating AI chat into WhatsApp and similar texting apps will make generative AI ā€œfeelā€ way more personal, and even intimate than ChatGPT presently does.

Is that bad?

It depends on your perspective. For instance, already, there are services that let dating app users use AI chats to ā€œpre screenā€ their matches, so they don't waste time texting someone who isnā€™t a good fit for them. If you don't like the sound of that, prepare yourself: youā€™re about to get very nostalgic for a pre ChatGPT era that is very, very quickly disappearing.

But texting with robots may have a fatal flaw.

NEWS FLASH: HAVING personal conversations with chatbots is fraught

Humans are social beings. Weā€™re all born with innate capacities for creating social connections. In fact, we need them in order to learn things, and for higher-order awareness and understanding too.

But chatbots aren't human. Theyā€™re not therapists. They are not friends, nor are they romantic partners. These are facts.

UNFORTUNATELY, THEY COLLIDE head on with another fact:

Even though people above a certain age may not fully agree, texting is social, and therefore very human.

Stay with me here.

If youā€™re texting ā€”which is a very human actā€” with a chatbot ā€”a very non human entityā€” some part of you, maybe really deep down, is primed for a human response.

Now if you don't quite get that human response, do you think that you might feel slightly, perhaps even just subconsciously, disappointed? And if so, how eager do you think youā€™ll be to return to that bot?

In other words, will that fact that generative AI chatbots feel almost human ā€”but not quite fully sapienā€” slow their adoption as people feel that somethingā€™s just off?

(A reminder that Iā€™m talking about personal chats with generative AI here. The eventual ubiquity of generative AI in professional contexts is clearly well on its way.)

THE SECRET TO interacting well with AI chatbots may well be the same as the secret to interacting well with people:

Manage your expectations.

Though not everyone agrees with this advice, many coaches tell clients not to have expectations of others. Expect too much from those around you, the thinking goes, and you're bound to be disappointed. If you donā€™t expect anything from anyone else, youā€™ll never be disappointed again.

It all lead me to this neat conclusion: If people expect human responses from AI, theyā€™ll be disappointed, and this disappointment may potentially slow uptake of the whole thing.

In other words, the mainstream product adoption rate of generative AI chatbots may be less dependent on the efficacy and quality of those chatbots, and more up to the expectations we quaint humans have of them.

Paradoxically, expecting less from AI, may lead to more use.

More:

Meta Q1 earnings were a ā€˜tour de forceā€™ Ā»Ā»

Mark Zuckerberg says Meta wants to ā€˜introduce AI agents to billions of peopleā€™ Ā»Ā»

How to get rid of Snapchatā€™s annoying, lying AI bot Ā»Ā»

We spoke to people who started using ChatGPT as their therapist Ā»Ā»

Do you suck at flirting on dating apps? This AI wants to help... Ā»Ā»

Written by Jon Kallus. Any feedback? Simply reply.

Reply

or to participate.