Fieldy AI Review - The $90 Wearable That “Remembers” Your Day / Limitless Acquired by Meta

Would you ever wear a device that actually remembers everything?

This is my take on the Fieldy AI.

Hi everyone.

This is part three in a series of physical AI products that I've been checking out.

The first one was the Even Reality's G1 smart glasses.

The second one was the Limitless AI pendant.

And the third is the Fieldy.

And this is my take of working and spending time with the Fieldy for over a week.

So in this series, I'll look at various aspects of the product.

I look at the hardware.

I look at the software.

And we'll also check out what integrations does it have.

And finally, answer the question.

Does it actually take a permanent place on my body?

And does it actually help me in my day to day life?

So this is actually an interesting product.

Full disclosure, this was sent to me by Fieldy.

They sent me the product itself and did not pay for it.

They also sent me a year worth of unlimited subscription.

But there's no other money that's actually changing hands.

And they don't even get to see this review before I publish it.

With that, let's dive in into the box.

So the box was pretty simple.

It feels like a very high quality cardboard, but I was disappointed when I opened it.

It only had the Fieldy pendant itself and just a necklace.

There was no cable.

There was no charging brick and there was no dock.

Honestly, I don't like that.

So many brands are actually cutting costs and not even giving me a simple cable.

It's USB-C.

So I would imagine that those are pretty cheap by now.

But having it within the box would have been a nice touch.

That's not there.

And I kind of miss it.

Anyway, talking about the hardware directly, let's look at it.

So this is a tiny piece of device comparing it in terms of size with a box of Tic Tac.

I think it's it's actually very small and light.

It has a USB-C port.

It has a button, which is clicky.

I also see a little hole for the microphone, a little Fieldy logo, which is engraved on one side and a little place for me to actually thread the necklace or any other piece of necklace, which I might have.

I don't see any other places for me to actually mount it.

There is no way for me to actually mount it to my clothing, to my shirt or wear it as a wristband.

Honestly, I don't like it just because I don't wear a necklace.

But maybe some people like it.

Reminds me of dog tags in terms of build quality.

The plastic is very rugged.

I've not been able to scratch it or bend it in the time I've actually had it.

It is very plasticky, feels like 3D printed, but it's actually not.

And it is not metal.

It doesn't have any magnets inside it either.

So you can't actually mount it with any other mechanism.

Of course, there could be a 3D printed clip on or 3D printed wristband, which you could do.

But I did not end up doing any of that for me personally.

I would just keep it in my pocket.

And whenever I was in a meeting, I would just place it on the table so that I could get the best sound which was available in the room.

The battery life was very good.

I actually charged it probably a few times and I would just leave it running.

It would run for 30, 40 hours and I wouldn't have to charge it again.

The battery was not even getting hot when I had it on.

It wasn't getting hot when I was charging it.

I would leave it charging all night long with my iPhone charger and it would just work.

In terms of the lights, there is one LED which actually shows you different statuses and different indicators on it.

And we'll talk about how that actually integrates.

And it's super confusing as to how it is.

But overall, I feel like the hardware is well built.

Maybe a smaller form factor would have been nicer.

A bigger microphone array would have been nicer.

I would imagine that would give me more audio, whether it's from me as the person or from the environment.

I couldn't validate what the quality of the audio is because they don't actually store any of the audios.

They only store the transcription.

Of course, every large language model has issues with transcription to a certain degree.

And that's why they ask you to add certain words which you feel like you pronounce a lot.

And the models tend to get wrong most often.

They are HIPAA compliant from day one.

So all of your conversations are actually safe and secure.

I would like many other companies to be the same where they can be GDPR and HIPAA compliant.

But I appreciate the fact that Feely was HIPAA compliant from day one and they were not storing any of the audio.

They intentionally choose to be more safe and secure.

Even though I would like to have my audio, but I understand their choice.

I think having just the transcripts is definitely the way to go.

So one other thing which I would like for it to have is Find My Support.

So it doesn't have any speakers, or it doesn't have any Find My Support or any other similar support.

I actually lost it one of these days, and I was trying to find it.

And the only way I could figure out how to find it was going to every single room, speak out the name of the room, and hope that this would still listen to what I was saying and then be able to transcribe in the app.

That's actually how I found it.

That was the only way it doesn't have any speakers.

It doesn't have any beeps, which it could make.

It would be a nice to have because it's so tiny.

I could easily lose it or misplace it.

So let's talk about the LED.

In terms of the LED, it has the option of having an always on LED light, which I definitely prefer because I want everyone to know when I'm recording.

But it also has different modes and different confusing modes.

I mean, if it's recording, it'll be a steady white light, as you see here.

If the battery is lower than 10 percent, it'll be steady red.

If it's muted or if it's running out of storage, it'll be a slow blinking white light.

And if it's not paired, it's blue.

I think those are the four primary ones which you need to worry about.

When I tap once on it, it should have a rapid blinking white, which shows that it is actually connected.

If I tap on it, it blinks three times.

I think no one can actually remember any of that.

It's so confusing.

And then it has different modes on charging versus not charging.

80 percent, lower than 80 percent, higher than 80 percent.

I think they've made a mess in terms of what the interaction is for the LEDs, and they could easily do a better job.

Yeah, honestly, I am not a fan of how they've actually implemented the LED.

So for the bottom, they only have support for one action right now, which is to start a conversation or start a memory.

Tap once, it starts a new memory.

And then it essentially categorizes this section of the conversation into a separate memory in my feed.

That's the hardware for you.

So let's jump into the software.

In my opinion, the star of the show for Fieldy is their app experience.

I think they've done a great job at building the app overall.

And I appreciate the level of care and detail that they actually put in.

So the app is split into three sections.

The first one being the memory speed, which is exactly where you would spend most of your time.

It automatically categorizes everything that you're saying, and it would group things based on the conversation that you're having.

The second is the task management and the task integration.

And the third one is the AI chat, which essentially merges the two to help me ask more questions about what's happening.

So looking at the memories, it's a simple feed, which is day by day, it's categorized by every single memory or conversation that you're having.

It uses in between silences to break down any conversation from another, you can go in deeper and have a summary of that conversation generator.

It also has different prompts and different templates, which are available at the click of a button to quickly switch between whether it's one on one, whether it's a lecture, whether it's one other, another type of conversation that you're having, and you could change the output.

It also allows you to create your own templates and own prompts so that if you have a certain kind of conversation that you have more often, you can actually tailor that prompt to be to your needs.

The other side of that same memory detail page is quickly being able to access the transcript.

So you can actually see the transcript quickly and be able to share it to other places, whether it's through messages or actually through to another LLM.

All right, going back to the memory detail page, we also see a section where they proactively actually ask you questions around, are there any other things that you would like to record around that conversation?

So it would surface some questions, which could be insightful, and you would want to answer those based on the memory that's in there.

The memory also has the option to actually log it against a particular calendar event, and it can also associate a memory for the location.

All right, once you're done with a single day worth of memories, it also gives you an overall summary of what's the overall day like.

And similarly, you have summaries for every single day going backwards.

You can swipe left and right to switch between days, or you can go to this entire list of all the days that you've had this running and recording, so that you can quickly jump between certain days, which you can identify quickly and go in into those memories to answer questions.

Going to tasks, this is super interesting.

So automatically, when it's recording all day, it would actually generate draft tasks for you to review.

So in my day-to-day, when my meeting is done, I would go in into the draft task section, and just swipe left or swipe right.

If I wanted to accept a task, I would swipe right.

If I did not want to accept that task in my task list, I would swipe left.

And that would quickly allow me to actually make sure that I have everything that I need for the work that I'm supposed to be doing.

It also has a list of tasks which I have accepted, and it automatically expires those tasks.

But it's interesting how that works.

They have tried to be sort of a bridge between a full-fledged task manager and a draft task taker.

I almost feel like that this is my capture device, which I use to capture all of my tasks, and then it could sink into different parts of your ecosystem.

Speaking of different parts of the ecosystem, we'll come to the chat at the end.

But before that, let's talk about the integration.

So they have a couple of integrations.

One is Google Calendar, Outlook Calendar, and Google Tasks.

So based on what we've seen so far, the memories can be associated to a Google Calendar event or an Outlook Calendar event.

Similarly, once you accept a task in your draft task and you swipe right, it can actually sink into your Google Tasks.

And I don't use Google Tasks often, but I was using it for this review, and I realized that whatever actions you take in Google Tasks, they don't sink back.

It's only a one-way sink to actually just create your tasks from PLD into Google Tasks, and that's it.

Whatever you do there, you do there.

And that could work for a lot of people.

I think I would have liked if it integrated with the Apple Reminders app just because I'm in the Apple ecosystem, but I feel like Google Tasks could also work.

Finally, they also have an API for me to integrate with.

So in terms of integration, that webhook allows me to actually get transcripts for every single memory that I'm recording.

It doesn't have a very detailed API right now.

It only has a very basic every-30-second push.

So if this is running, it'll keep on pushing the transcript every 30 seconds to my webhook, and right now I have it set up with PipeTream and pushing into my Roam instance.

But technically, you could have your make.com or other webhook implementations, which are simple, no-code implementations, and that could get you all the transcript you need every 30 seconds.

All right, going to the chat.

So just like every other AI chat that you've seen, this most likely uses ChatGPT for its chat, and you can go into any memory and start a conversation.

The interesting way Fieldy has implemented is that it's a very simple mechanism.

It's a single chat thread, which you can continue to have conversation on.

It doesn't allow you to switch, and most likely, once you run out of context, that's when it goes on to a new chat.

But right now, there's no way for you to manage different chat instances.

The chat also has a way for me to record my audio and then ask a question.

It doesn't have a way for me to respond back with voice.

I feel like whatever Fieldy lacks in terms of hardware, they actually make up in terms of software, and the software is continuously improving.

I like the fact that they're focused on both ends, updating the firmware and also updating the app experience.

They don't have a Mac app, unfortunately, but you can run the Fieldy app, which is the iOS app, if you have an M-series Mac with you.

So that just runs in the phone sort of mode, but it would have been nice to have a Mac app also in terms of software.

Would I say that this is close to being an assistant?

I don't think so.

I think it's still a recorder, even though it has some sort of offline recording capabilities.

It has a way for me to create tasks, but I wouldn't call this an assistant yet.

Yes, it can identify certain tasks, but it can't do those tasks for me.

It can only assign them back to me.

It can't even assign them to anyone in my team.

I would not classify this as an assistant yet.

Maybe an intern, an early stage beta intern, that's what maybe this is, but it actually doesn't do any work.

It only does that first level of work of creating the tasks and maybe categorizing my meetings, but it doesn't really do anything else.

I think having other integrations with my project management systems or having those tasks move on to my teams might actually be more interesting than just Google Tasks or any other simple task management integration.

So who should consider this?

I think anyone who's having a lot of meetings every day, founders journalists anyone who spends a lot of time speaking would actually like this because this would allow them to understand how their day has been like and ask questions and recall things.

So recallability is easier with this.

If you're just doing your work in front of a computer, you don't necessarily need this.

Think of it as going back to those traditional recorders which Sony used to make, and you would send that out for transcription to someone else, whether it's a lawyer or whether it's a founder or whether it's a journalist, they would carry those around.

And this could easily replace something like that.

It gives you instant transcription.

I think teenage engineering also makes a recorder and none of those devices actually fit you in your day-to-day like this.

The only caveat, you have to wear it as a necklace.

You can't wear it as a wristband or a clip-on.

So how much does it cost?

The hardware is actually very well-priced.

It's actually only 90 bucks.

And then there are different plans which are available for you.

There's a free plan, there's a plus plan, and then there's a pro plan.

The free plan actually offers 150 minutes every month, and which should be more than enough for anyone who is infrequently taking meetings.

So you only spend the 90 bucks on the device.

It doesn't give you the option to actually export transcripts, but that's totally fine.

Once you go to the plus plan, you get a lot more minutes.

It's, I think, 10 bucks a month for an annual subscription or 13 if you're doing monthly.

And the unlimited gives you the 20 bucks a month subscription to you.

It's the same as all other AI model subscriptions like Chad GPT or Claude, they're all at 20 bucks a month.

But that's where they're priced at.

I think they're a little bit more expensive, but I understand that they have to pay for the transcription service also, just because that's the added cost if you're running this all the time.

So does it dethrone the Limitless Pendant for me?

Well, honestly, I can't make up my mind.

I think it has its own value.

The app experience is nice.

And for the time being, I'll actually be using both the Limitless Pendant and the Fieldy AI.

I think there's a lot of value in exploring what different companies are doing with these products.

I like the Fieldy app experience a lot.

They've hit a home run.

I would prefer to sort of learn more from what everyone's doing and try and see what integrates better in my life.

It's not like the iPhone or the Android phone, which is sort of the de facto standard in the industry.

We are in an age where everyone's trying to make new devices, whether it's the OMI, whether it's Limitless, whether it's OpenAI.

And I think I'm here for it.

I want to explore what's happening in the ecosystem and pick the best device, which actually fits for my daily workflows.

But for now, I'm actually just using all of them.

All right, if you like what you see, please subscribe and leave a comment.

I'll see you next time.

I'll see you next time.