haxplorer

(I've been tracking this product's development over the last 6 months and have been excited enough to post about them. Here is a blurb about the product's founder - Jose's - own words)

TL;DR: We're building the first-of-its-kind Live Video Agent that can handle sales and onboarding calls. Imagine a sales rep or onboarding specialist on Zoom—our agent can do it all.

What makes our agent unique is its ability to log into software and perform human-like demos, showcasing how the software works. Our agent is trained on the UI to present every use case based on the customer's needs.

We train our agent using videos of sales and onboarding calls, giving us two major advantages: any previously asked question can be answered by the agent, and onboarding is seamless and less time-consuming compared to human sales reps.

We launched on Product Hunt a couple of weeks ago, and our SDR agent made 143 calls, securing 8 meetings, which converted into 3 customers. This is our proof of concept—if an unknown brand can acquire customers with an AI agent, imagine what an established brand can achieve.

We're just starting and would love to get feedback from anyone interested.

Founders calendar - https://cal.com/fluid/30min

Not really funny. When I ask it to perform regular tasks, it does it with no sarcasm or dark shades in the way it is performed. Not really the evil twin of ChatGPT

We've found some of the best engineers from the open source communities, but it's never been an easy task. Outbound on this has always been a lot of work, with low response rates.

Good to see a two sided solution for this. Keen to try out soon

While their current focus is on badges and rewards using tokens, I see other possible directions for this product that could transform the opensource landscape. The founders seem to be looking at this direction in the mid-to-long term as per a comment on ProductHunt.

The early contributors to an Open source project often go unrewarded when it becomes a commercial hit.

Cases in point - Mongo, Hadoop (Hortonworks), Memsql, etc. started out with a good amount of community support, and later went on to becoming commercial successes. While they recognized a lot of the contributors through various programs, and sometimes hired them in, the others who didn't join their journey full time might not have benefited.

Many other companies start with a "Develop in the open" philosophy now, despite being VC backed from early stages - cases in point Meteor, Gitlab, Appsmith, etc. The community engagement is slightly on the lower side once the project is well staffed. The motivation seems to go down.

What if the contributors could have a pie of the company? A very small pie that is proportional to the importance of the feature or bug fix at that point in time. But, as the company grows, the value of the pie grows. Just like an ESOP, or any token. Could say a project set aside 20% of their enterprise value as tokens, and use them as rewards to drive their roadmap with the community? Higher the importance, higher the bounty. That could change the dynamics around OSS contributions, which are currently driven more by the interest of the contributor more than the importance of a feature or a fix to the project.

I see Aviyel could also potentially drive such a shift.

Saw them on Producthunt today - https://www.producthunt.com/posts/aviyel

The comment referred to at the top of this post - https://www.producthunt.com/posts/aviyel?comment=1785981

Very helpful list. But it would be good to see a list that is organized around objectives than the current set of categories. Ultimately, one asks these questions to get a certain outcome or information. Starting from there would be better.

We use a mix of confluence and slack groups. We have one slack group per type of learning (RCAs for failures, best practices in engineering / manufacturing / logistics, etc., Top of the mind thoughts, interesting reads). We use confluence to do detailed documentation or writing, and slack to share and drive visibility.

Cant expect more transparency. Thanks for reasoning out.

Another handy feature would be - ability to call teammates from within hipchat. When I read the title, that is the feature I expected. But even this is pretty neat.

A fixed price would give peace of mind to someone like me. A pay as you go plan would just increase the uncertainty. I would always choose a fixed price plan, unless the cost saving with the pay as you go plan would be significantly lower given the average number of chats that would happen on my site.

Good effort on the mocks, all for just an April fools' day :)

Didn't know Reelsurfer lets you create >1min clips. Is that a pro feature? It always restricted me to a few seconds.

I'm Raj from Markupwand team.

You are right; the code that Markupwand currently gives isn't upto the mark. It makes a few wrong guesses on how the PSD should be interpreted. But that is just for a few more days.

We are coming up with a semi-automatic version, where Markupwand would take inputs from the developers on how the code should be structured. That would make the code usable.

I think what he means to convey there is that a person with a low predisposition would be affected by the actual content in the drug than just taking any pill.

Including predisposition to placebo as a variable would just increase the number of experiments needed. Instead, if you could eliminate a variable from the system by preselecting for people for whom the variable doesn't apply, it would reduce the time to experiment

It is a very healthy business model. Since repeated transactions are kind of guaranteed to happen, the company could spend confidently on acquiring new customers. They don't have to spend time and again on the same customer to acquire newer transactions from them.

Here is how most non-subscription based e-commerce companies today work: The company shows some traction, gets good funding, acquires customers at a cost higher than the value of the current transaction, with the hope that the customer would later come back to buy more, and the company could recover the acquisition cost over time, and make profits. But every other competitor in the space is doing the same, and the customers have multiple places to buy from. So, the company ends up acquiring the same customer time and again, and all these costs get added up to the customer acquisition cost. But in the early stages, the lifetime value of the customer wouldn't be evident, and the company might end up spending lot more than the lifetime value to acquire transactions in the short term.

This problem doesn't happen with subscription based e-commerce companies, as long as they provide a good service and are able to retain the subscriptions. They could pay for advertisements on a cost per acquisition model, and count only new customer acquisitions, and subscription upgrades are conversions. This way their customer acquisition cost would be constant and under control.

Here are some of the areas where subscription based e-commerce would work well: 1) Groceries, fruits and vegetables 2) Cosmetics and sanitary products - Facial creams, lipsticks, skin care products, after shave lotion, shaving foam, kitchen napkins, toilet papers, sanitary napkins & tampons, condoms, cleaning agents, detergents, etc. 3) Baby products - diapers, baby oil, talc, etc. 4) Medicines - Diabetic, blood pressure control, cancer drugs, etc. - But there are regulatory concerns here 5) Undergarments, socks, etc.

Most of this is already being done by some company or the other.

I've been through this precise problem both from a developer's as well as a consumer's standpoint.

As a developer, when we built Unifyphotos(http://export.goyaka.com/) to transfer photos from flickr to facebook, and got some traffic peaks, it was lot of work to scale up the download_from_flickr -> scale_photos -> upload_to_Facebook process, and we had 6 machines just doing this. Wish Filepicker.io was there back then. We could just have worried about the features and not the infrastructure. Even better, if Facebook used Filepicker, there would have been no need for a service to migrate those photos.

From a consumer side, I've often had to download my Facebook profile pic and upload as profile pic on sites like trello, basecamp, etc. If only they used Filepicker.io...

Wonder why Olark didn't build it. I just didn't realize that this was missing.

It is open for everyone now. You could sign up on the website

http://www.markupwand.com/signup

FPComplete seems to be a Haskell services firm. Not sure how big a market that is. But few companies use Haskell for Business Intelligence. May be they are looking at them.

Sorry about the fifth point not being clear. Will try to make it clear.

Some designers produce effects using indirect methods instead of using the photoshop effects. These should be avoided as much as possible, for the code to be clean.

We don't intend to open source it

I haven't used Youtube ads personally. But I've seen how and when companies typically use them.

Mid sized companies do 1. Retargeting 2. Demographic based ads(especially location based)

Big companies do branding.

Retargeting: If a user visits your website, you could ask google to add a tracking cookie to that visitor. You add additional information about that user like which stage of the funnel the user was in(eg., visited a product detail page, visited pricing page, went into checkout and dropped out, etc.). Based on this you could target a specific version of ad to that user, wherever he goes.

Most ads I see on Youtube are from companies that are in growth phase doing retargeting. Eg., Legalzoom, Zappos(and a few other e-com companies)

Demographic targeted ads: I saw that the set of ads I'm being shown have been changing drastically based on my location. And the ads I see also are different for me vs my friends who are slightly older. They mostly get insurance ads while I get beer ads.

To re-iterate, all this is empirical observation, and I don't have a first hand experience about this. But hope this helped