cmikec

Magic | http://getmagicnow.com/ | San Francisco, CA | ONSITE | Full-stack & Back-end Engineers

Magic is building the future of human-computer interaction: natural language chat that is capable of handling all possible human requests. Our users interact with us 24/7 over SMS. We have no user-facing software. Our back-end is a hybrid of human and machine intelligence. We are growing rapidly and our product is in high demand.

We are looking for engineers that understand practical business problems and can deploy code to production on a rapid cycle and see measurable impact. Magic’s engineering task is to use software as a weapon to tame a staggering level of real-world complexity. We test our code on live firehose of real data every day.

We stay up late coding and we love it. We like to use the latest and greatest technologies but we also know when a quick bash script is the right solution. We are highly iterative. We are relentless. We don’t stop until the job is done.

To apply, email [email protected].

We are highly focused on software that automates and/or makes parts of Magic and Magic+ more efficient. Most of the founding team are engineers, including myself. At this time we can't say much about our specific technical approach, other than that it is a very difficult problem, we have a novel approach, and we are making progress.

Before we rolled out Magic+ today, we did quite a number of split tests regarding prices and messaging. We also have had a private beta of Magic+ going for some time. Venturebeat as well as a few others seem to have screenshotted our website when it was in various stages of being split-tested.

Yes, you can use the service from other countries. Magic+ is actually really convenient when you're traveling. We have phone numbers that work internationally.

User retention and product quality.

If we aren't efficient and cost-effective, we won't be able to retain users and our product experience will be severely harmed. Whether we are charging hourly or not, it must be the case that the value we offer the user is greater than or equal to the cost that the user is paying.

Needless to say, at Magic we hold ourselves to a very high standard when it comes to product quality and value. And, we use the product ourselves and we pay hourly.

Bitcoin was nowhere near as popular of a payment option for Magic as some people expected it to be. For simplicity's sake, we did not include it right now in the Magic+ service. If you would like to use Magic+ via Bitcoin, please text in and mention that fact, and we will certainly accommodate.

Thanks for bringing this up. I've addressed similar comments elsewhere in this thread so I'll try not to repeat myself much.

- We really feel terrible that some early Magic users had a less-than-magical experience after our initial, spontaneous launch. I'm really sorry about that. I'm glad that you got your refund. If you're interested, reach out to us and we can probably track down exactly what happened and what broke down in the chain of operations.

- This is part of the reason that we created Magic+. We wanted to be able to hold a very high bar for top-tier service and guarantee that you would have it at all times, no matter what.

- We'd love for you to give Magic+ a try and compare it to your experiences with Magic. I'm highly confident that you will feel a major difference, however I understand if you don't feel inclined to do so after your past experience.

Hi, this is Mike, one of the founders of Magic. Really sorry to hear about your experience here. Did you get a refund? If not, please text in to 83489 and mention this post, and we'll be sure to get it to you.

I can absolutely admit that when Magic first launched, we were over capacity and a certain percentage of orders went awry. I hate that this happened, but we really did not predict that the website would go viral so quickly. In fact, today, we still have such high demand for the service that there is a long waitlist for the regular service, and our $100/hr service is getting a lot of signups very quickly as well.

This is part of the reason that we created Magic+. We wanted to make sure that people who had a very high bar for service quality would be able to use the service right away and be guaranteed to have the absolute highest level of service, no matter what.

I completely understand that given your past experience with Magic, you might not want to try Magic+. However, I can confidently say that if you try it, you will find that it feels very different from the Magic that you tried. Your keg will be there.

Thank you so much for your comments, feedback, and kind words!

This would be a real risk if you hired an ordinary personal assistant. This is actually exactly the sort of challenge that traditional assistants pose and that Magic+ is designed to fix.

Magic+ is actually not just one person, but it's a software-driven service with a highly trained team of professional assistants that are held to a very high bar. They are managed, supervised, and trained by us so that you don't have to worry about it.

And, on top of that, every task and request is managed from beginning to end using custom software that is designed specifically to monitor the execution strategy and enhance both the quality and the reliability of the output. It's difficult for me to say more right now about how this works without disclosing IP that we cannot disclose at this time, but I hope to be able to say more about how this works in the future.

A good way to think about Magic+ is that it's a natural-language command-line interface for complex tasks that require some human interaction. A component of how we handle your email is more like a ~/.procmailrc file that can pipe things to a highly-supervised human assistant than it is like having a random person somewhere in the US logged in to your Gmail. You tell us how you need things done, using natural language, and it gets done, assuming it's possible.

Again, I recognize that this explanation is not as detailed as it could be, but I hope that it helps. I will think about how we can explain this better before we are able to disclose the inner-workings of our software.

I'm really sorry that your initial experience with Magic last year wasn't ideal. We have come a long way since those initial days. Believe it or not, we actually did not plan that original launch -- rather, we released the website and number to a few people and it went viral practically overnight. We set up a waitlist as quickly as we could, but there definitely were some people in the beginning who did not have a good user experience. I hope that if you asked for a refund back then because you were dissatisfied, that you got one. If you didn't, please feel free to text in now and claim it.

Some of the themes you're bringing up here actually point to why we created Magic+ in the first place. Magic+ was developed as a response to the needs of the most active and demanding users of our standard Magic service. In particular, a common demand was to hold a very high bar for quality and seamlessness of experience.

Right now, some of these user-requested experiences are more expensive to deliver than the original Magic pricing model accounted for. Our plan is to offer Magic+ at this price now, because there is good demand for it at this price, and drop the price over time to reach a larger and larger market. $100 dollars an hour is a lot, but our most active users tell us that for this level of elite service, it's a bargain.

Thanks for bringing this up. We've adjusted these bullet points to make them more accurate.

These are all real requests that our users have actually used Magic+ for. We obviously cant divulge their identities, or show photographs, etc. (since these are personal things), but it's important to convey how people really use the service. I'd love ideas on how we can communicate this point better. Part of what our users love about Magic+ is that you truly can do some pretty amazing things with it.

Currently these reports are sent via email. We may have an app at some point in the future, but currently SMS and email are more than sufficient.

Many of our users give us access to various aspects of their personal and work lives in order to gain the benefit of efficiency and saved time. We also set up users who don't want to share their primary email with email addresses at our domain.

Thanks for the feedback. You're right, you bring up a good point that isn't explained as much on the website as it could be. We actually track time down to the second and display really detailed reports to users every day about how the time was utilized. Quite often a simple request can end up being $3-$4.

Thanks for the feedback. Magic has been in an interesting position as a startup ever since our website went viral in Feb. 2015. Starting then, we've had more demand than supply for our product every single day. That's why we never iterated our original landing page -- marketing was never our biggest problem, in fact it was our smallest. That said, I completely agree with you, that there would be much more powerful ways of communicating that you can trust us. Most of our signups come from word of mouth from real users at this point though, so it's not as much of an issue.

For our $100/hr Magic+ product, we display detailed reports daily to users that show exactly how time was used down to the second. I agree that communicating this correctly is key. If you're interested in trying it, I'd love to hear your feedback on how we display this information.

Yeah, we've been receiving a lot of requests like this. Right now we are doing our best to handle them all because we believe that happy users will be purchasing customers. We'll have to see how it goes.

We noticed that people were more confused by the ambiguity of the nondescript fee, than us just saying that we'd tell you the total amount. You are right though, we do need to figure out a better way to explain this. If anyone has any ideas, please let me know.

This was pretty funny, actually. They saw the old version I mentioned above. I was surprised when someone posted it on there and I was actually glad that the reaction was lukewarm, because we didn't think we were ready to launch.

And then...boom.

Wow, everyone! This is truly amazing. We released Magic under 48 hours ago as an experiment and a side project, and now the traffic and requests are streaming in faster than we can handle.

Don't worry - we have a plan to handle it. For now we are closing down free registrations so that we can focus on delivering the product to the awesome people who have signed up so far and who are using the service.

I've replaced the phone number on the page you see here with an email opt-in waiting list where you can sign up to be notified when Magic is available to you.

In the meantime, I have added a Stripe button for $20 after the email opt-in where you can gain access now if you want to get in right now.

Thanks everyone!