I brought up Poke a few days ago, how it's "being my life support."
I'm not exaggerating. When I posted the "super obnoxious" and "above-everyone-at-amherst" post about hiring assistants, I didn't know if I just waited for a few days (which wouldn't have happened because at that moment I was extremely busy), I would've my idea of hiring a human assistant completely.
The Interaction Company of California created the greatest form factor of personal superintelligence. The concept they have is how I'd define personal superintelligence.
So what is Poke? Poke is a chatbot that lives on iMessage, SMS, and Whatsapp. It reminds and helps manage your emails and calendar, along with some third-party platforms like Notion. If we look at how Mark Zuckerberg defines personal superintelligence, this is basically it. It's just that Interaction isn't building its own model.
The product is still flawed. And I pointed them out on X as well as emailing my complaints to the dev team, which in return, they not only apologized but also gave me a custom $50 promo code with my name on it so I can grab their merch for free. I got a t-shirt and poster from them btw.
Anyways, here are the problems I've encountered so far:
- Daily schedule/email briefings are not there even when I've set up the proper automation for it.
- It can compose email drafts, but is unable to save them in the drafts folder.
- It can’t access the internet to provide daily news briefings from time to time, and when it does, it would sometimes pull out news from months ago, which is older than just “yesterday.”
- It can’t access Notion even after multiple reconnection attempts. And even when connection is successful, it couldn't read all the variables in the databases correctly. These including item "status" and "date/deadline."
- It hallucinates when checking my emails and calendars. For example, it would sometimes make up things that the other side or I didn’t talk about in an email thread, or it will show me a fake calendar schedule when I ask what my schedule is on a future date.
I think information accuracy is the most critical issue. If it keeps hallucinating when I ask questions about an email conversation I had with someone weeks ago and it gets it wrong, am I supposed to go back and check the email myself to verify, which results in me wasting my time, or should I trust the wrong information? Imo, this needs to be the TOP priority for the Interaction team, and I hope they address it correctly.
But besides that, it does its job 90% of the time.
It has context about what I'm working on. It does a better job than the human assistant I hired. And it works FAST. And it's the best form factor anyone could ask for.
That is my honest review.
But how would this impact the human workforce?
For entrepreneurs or people who are high-level executives, this thing will be very handy. But for people who are just serving other people, like assistants or secretaries? They will be replaced.
When I told a friend about how I started considering replacing my human assistant with Poke, which does its job even better for less of a price ($10/month instead of $200/month), this is what he said to me:
when did poke come into existence
also, even if poke came from yesterday
idc
it's still a fucking dick move to hire and immediately fire someone for no fault of their own
Idk if I'm being reasonable or being overly pragmatic, but I'm not really seeing his point. These situations are inevitable in the future. If everyone starts relying on robots for farming, manufacturing, or transportation, are we really going to spend every minute sympathizing with the people who lost their low-level jobs but are still getting subsidized by the government?
Something to think about.