Former meta prick, sorry employee:

> Meta's strategy

Meta has no strategy. The people that made llama work left two years ago. MSI are a bunch of charlatans who are trying to change the culutre and tooling of meta from the outside (I can see their point, trying to work outside of the defaults is incredibly hard.)

its basically a bun fight, with all of your productive workers being poached by frontier labs, and everyone else is either a new grad, burnt out, or a master bullshitter.

I think you're right, but I'd add that... if this wasn't obvious to you without being told, you should really question your worldview a bit.

A lot of people think that chat.com is an AI called "chat". Most people don't know acronym "LLM". Most people don't know about the existence of companies called "OpenAI" or "Anthropic", or that they are worth a lot of money. Etc.

At $dayjob, a company with ~10k employees, there are like 3 people who cannot shut up discussing the latest models in the internal AI forum, 10 others that constantly compare models and talk about "strategy" of using the models.

The rest of the users just use whatever the default is and get their work done.

I'm confident that Muse can do lots of agentic tasks succesfully for normies on outdated models. Like buying sneakers, booking appointments, dealing with government forms.

I think of my mom who doesn't care and doesn't understand the difference between google AI Overview vs actual google search results. "I looked it up" - she doesn't care what tool she is using she just wants answers to how long bake pork or the best throw pillows to buy or whatever else she is looking up.

She's mostly using google AI tools because it's easy and free and on her landing page.

> So, us serial HN users and commenters are the extreme minority

Words to live by.

I'm a coder and a daily driver of ChatGPT and I don't really know what Sol is or how I can access it. I follow discussions here and know of its excistence, but I can't remember its relation to other models or tiers by heart. What ever the default thing the website is giving me has been good enough for my needs for years.

Given that cost of building is down, strategy of most companies is turning into "do everything" and see what sticks

I confess that when I ride the subway, I like to peer at other people's phones to see what they are doing. The ones that interest me the most are other people talking with an LLM. I am always surprised by the wide variety of people who are using LLMs to ask an infinite variety of questions. From the surface-level PR annoucement here by Meta, this seems like a good business idea.

> "just ChatGPT, what is Sol?."

i'd really like to see some analytics on what proportion of users have ever switched the model in chatgpt because I anecdotally believe it is < 5%

I think we are lacking in the easy to use side of things. Once you go beyond browser based chatbots things become hard to set up.

I see lots of posting about using AI to create patches to get new apps working in wine, or reverse engineering firmware. But every time I’ve tried it it’s been quite a lot of work to set everything up. Even more so if you want VMs and some level of isolation from your sensitive data.

Having something that’s just a one click product that just works is more important than the smartest model.

They are not 'normies' - we're just 'nerds'.

You're driving a home kit car, they're driving something out of the factory.

Eventually - we'll probably drive factory variations as well.

> I think Meta's strategy is to capture the 'normie-tier' of AI users.

For what it's worth, maybe that would help people who struggle with technology in general? Because as far as UI/UX goes, a chat interface for an agent is about as easy as it gets (not every site having its own crappy assistant that can't really do anything useful for you, but is slapped on there for the sake of a checkbox).

I watched the marketing video and to me it seemed pretty nice and practical for the average person.

in a way it's not much different from a time where the E icon of internet explorer was considered "the internet".

It seems like on the “Chat” mode of ChatGPT it doesn’t allow switching the model. It’s just “5.6” with the effort level

I use claude all day every day, and sometimes codex, and I still don't change models. I guess I'm more interested in accomplishing the thing than in experimenting with a dozen ways of accomplishing the same thing.

To be fair I stopped caring about specific versions as well. I don't seem to notice the big difference anymore on daily tasks and questions.

Yesterday I helped a maaaaybe 40 year old woman connect to the wifi at my local library so she could get on a zoom call. Made me realize most of the world is exactly like that.

I agree, and stay away from meta generally in life wherever possible

That said, I've got more time for projects this week so flew through my Claude max plan. Now I'm trying muse spark 1.3 with opencode which is free with pretty generous usage allowance. First time I've been impressed by meta's AI products. It feels not far below opus 4.8ish

Even with all of this nuance why would the regular joe go to muse when they have chatgpt and gemini? Most of their actual productive data is likely with Google anyway.

I think all AI companies want to capture normie users too, it's a huge market.

All ChatGPT, Claude, Google, Microsoft have/want to have their own assistants.

This is not limited to non technical users. Plenty of people don't care to switch between models and just pick the latest model.

I met a normie who was using ChatGPT to write production code… in instant mode. No thinking enabled.

Literally what I was about to comment. This is meant to capture the normies, the busy parents managing school and work, the grandparents who want to make most of their retirement.

I am trying it now. It let me name my muse as Satan. I appreciate that it understands my humor, but I'm also not joking at the same time.

Yeah this to me is why Apple wins

Many people deep into tech (like most commenters here) may fall into the trap of overestimating average user's knowledge or interest in the details of the tech: https://www.explainxkcd.com/wiki/index.php/2501:_Average_Fam...

It's not a bad market to aim for. No one else really is.

Cant wait for Snitch Bench results of this

> I then realized

Oh, your grandma doesn't know which android version her phone is running? Or the build number of the Windows on her laptop? You just realized this?

> I know most of us here on HN track model releases quite frequently and discuss every parameter weight out of them

You know it? I sincerely hope you don’t and that it’s not really “most” of us. That would be incredibly sad. Tracking model releases and discussing “every parameter weight” is very far from intellectually curious discussion (what HN is ostensibly about).

> I then realized that most people just stick with whatever default they're provided with (…)

> Curse of knowledge and all.

If there’s one thing many of us on HN suffer from, however, is intense hubris. The juxtaposition of claiming “curse of knowledge” while admitting to not having known the most elementary principle of designing for people is a tad tone deaf. Yes, of course most people just stick with the defaults, that’s why dark patterns work and hiding settings is a thing. The industry has been using those strategies to manipulate people for years and years.

To me it sounds like they are simply taking social moon shots. They tried to do a phone, tried to make a kitchen videophone, tried to make their own cryptocurrency, and tried to dominate open source AI, and now this is trying to make something like a digital robotic assistant. It might have reached a peak with the metaverse, when Zuck thought his corporation was destined to be the ruler of a Snow Crash style metaverse, to the point of renaming it. Maybe that's still the plan and they are looking for something that will lead to another metaverse play.

[dead]