Product Trends I am seeing in 2023

Hey guys,

Found some potential trends based on my exploring different communities.

Feel free to add anything else you guys are seeing or if your product falls under one of those!

Notionfying - The .so trend

Notion has a very unique UI/UX that makes it really easy to put thoughts on screen in a nice minimalistic format. I'm seeing products like Tally.so that make form building easier or Cirlce.so for community engagement. What else can we make easier for users to make and look at?

“Quick links”

Making an event can be tedious, especially with half of your friends not using Facebook anymore, hence features like FB Events are friction-full. Products like Partiful and Envited 10x making events. Other products are vidday which you can use to quickly make collaborative videos.

MicroSaas

I think this one has been mentioned a lot.

Platform-specific

GummySearch for Reddit or TweetHunter for Twitter. Platforms, where there are a lot of large communities, can be leveraged to grow your product/brand, find insights, etc. What other platforms can benefit from tools like that? Or what new APIs are opening to the public that would enable this?

Most obvious one - GPT Wrappers (AI products)

If you played around with the APIs or even GPT Chat you can tell how easy it is to deliver massive value on something like content creation or summarization. The key for AI products will be finding niche applications, fine-tuning, and figuring out creative distribution.

Anyhow this is a conversation, would love to see what you guys think. Follow me on Twitter, I might post useful stuff time to time https://twitter.com/mikeshaimerden

Michael Shaimerden

on December 14, 2022

Say something nice to michaelrecycle…

Post Comment

2

AI stuff is definitely getting more popular. And I think that although GPT wrappers are popular, there is a push from some people to move to other AI solutions that are not OpenAI.

For example, the move towards more open source AI solutions.

If anyone is curious, I'm currently developing my own AI API that you can easily build wrappers out of. It'll host an open source model called BLOOM that has less restrictions than GPT-3 and will also be cheaper despite being a similar size.

Feel free to follow me on twitter to tag along on my journey if you're interested: https://twitter.com/therealetch

RichardGao

·
4 years ago
·
Reply
1. 1
  @RichardGao - I am been trying to make sense of Evoke as a business person. I get this nudge it is likely very helpful but cannot twig.
  
  I have watched your hitting many posts here on IH, plugging Evoke. I never like to see such nonsense -- as it is not what IH about. IH is about helping one another honestly. Perhaps your from an old sales background, and cannot get the salesman out of the indie....
  
  Perhaps you are missing a key point, perhaps it is an opportunity. Where is the evoke directory of APIs supported, and what each API affords from a business perspective?
  Like this directory of replicate AI models https://www.replicatecodex.com/, but for API or even open source AI.
  
  Are there any GIFs to show me simplicity of using Evoke. Solving x y z business problem ???
  
  Open source is absolutely king for the SMEs I work with. But paying for open source seems counter-intuitive.
  
  We all use our GPT4, chatgpt plus subscription so dont need any reinventing the world. In fact when we see dev do that, tells us, they dont get business.
  
  open data sets is still in immature phase but that is the trend
  
  open dashboards to drive customer and citizen awareness and actions is trending too
  
  the noise on the web is typically too much for most SMEs. And every little piece costs another subscription. This is a major turnoff for SMEs, just trying to wrap their heads around things and play. SMEs more than happy to pay for value, but not for playing and gtk.
  
  FYI: I did signup to the discord but that is overwhelming for a non-discord user who prefers distraction free, noise free environment.
  catalystNewbie
  
  ·
  4 years ago
  ·
  Reply
2. 1
  
  I'm bullish on other AI solutions but for it to be as good as openai they need a lot of capital, not all teams have that right now.
  
  Michael Shaimerden
  
  ·
  4 years ago
  ·
  Reply
  1. 1
    
    Oh yea definitely. We're not going in the same direction as OpenAI. We're just making open source models more accessible by hosting them on the cloud and granting API access.
    
    We'll have other AIs on there in the future as well. For example, Whisper (subtitle and translation) and YOLO (img recognition).
    
    RichardGao
    
    ·
    4 years ago
    ·
    Reply
    1. 1
      
      Ah right, we use whisper right now. So you help outsource the CPU utilization by making those tasks endpoints. How much do you charge?
      
      Michael Shaimerden
      
      ·
      4 years ago
      ·
      Reply
      1. 1
        
        Yep, we help outsource the CPU/GPU stuff.
        
        We charge based on compute time used, as we're still WIP, we haven't decided yet. Generally, the more computationally advanced the task, the more compute time it will take and the more it will cost.
        
        For example, generating an 8k hyperdetailed image would probably cost more than translating a few sentences in Spanish, and Whisper seems to be relatively low on compute usage, so once we get it on the cloud, it should be pretty cheap to run.
        
        RichardGao
        
        ·
        4 years ago
        ·
        Reply
        
        1
        
        The thing about Whisper isn't the CPU cost, it's the fact that you'll have to have a lot of multiprocessing going on per one request, let alone if multiple people try to call, you'll then have to have some kind of a queue system for users, so centralizing CPU heavy tasks unless you are AWS might be too expensive and hard to scale
        
        Michael Shaimerden
        
        ·
        4 years ago
        ·
        Reply
        
        1
        
        Interesting. We'll see how much there is in terms of demand for whisper, and we'll integrate it in the future.
        
        I'll post updates on my twitter.
        
        Thanks for the information!
        
        RichardGao
        
        ·
        4 years ago
        ·
        Reply
1
Biggest trends fm, are 3:
1. Mobile capture to transactional systems. Crowd source transparency dashboards.
- Sure apple's live text API means iphone apps can go first. I will be surprised if android does not offer a similar powerful AI api. This technical capability enables a host of apps that eliminate the tedious task of data entry.
- Broad range --- From Citizen Scientist apps like The Ocean Cleanup uses to report local issues (trash in rivers etc..), thru to 1-click list my item on eBay/TradeMe, thru to Citizen report apps of local issues to Council (blocked drain, empty dog poo bags, trash, missing sign, illogical road work etc... ) and to Mobile Shopping.
- This shift will enable citizens to be the point of capture versus business (or customer of business). That is a signficant shift in our world.
- All these mobile capture apps can produce data-sets (many can be open) and boost the shift to crowd source transparency thru dashboards. This will serve to counter the much marketing nonsense by the big companies. Citizens love evidence. Citizens love transparency. (a shift in power-dynamics)
1. Mobile shopping
- Mobile shopping has started. However the app that enables mobile shopping independent of all the major grocery stores will be the winner.
- Sure here in NZ, the local stores will launch their own app but I would prefer 1 app I can use in any grocery store. It means the power of online shopping is now with me, at my finger tips.
- How it could work -- single scan, continuous scan options are needed (as with the sparkScan SDK), so at the checkout (POS), I simply QR code my shopping cart in to the grocery store system.
- Benefit -- No more standing at the checking scanning my items. No more getting ripped off on prices. The mobile shopping app allows me to plug in my 5 local grocery stores and it shows me the online price comparison as I am walking around the store. Often the in store price differs to the online price for the exact same store, in my country (NZ). Downright monopolistic behaviours by our major grocery chains need to go. Same in Australia too.
1. No more need to pay University/College fees. Perhaps even school
- Likely seriously, I have never used either of my degrees. Waste of money. Huge debt. I learn by doing. So a shift to apprentice led schooling/higher education will start. Like the trade sector has never lost, will be applied to white-collar working fields.
- Self-learning, is now very possible. Grab a course outline, grab the course outcomes and gather the learning materials. Then go for the test. Sure we can still pay for the test/certification certificate. Sure we can even pay for mentors or coaches. All this is value add. Receiving knowledge is no longer value add. It is readily available and for free.
- Birth of challenge hubs to guide app development, solution development. Businesses post a challenge (aka something they really need to solve a real problem). Post a bucket of cash and citizens submit. It is already happening in the waste-free world, circular economy world.
- Now with GPT4 this can be taken to entirely new level. Chatgpt GPT4 gives me clear instructions, step by step, for anything I ask. e.g. How to build an iphone app to do xxx
- The winners in this sector (bigger than a niche), will have the AI model text to video that takes a line of instruction (e.g. code) and creates an in context how to video.
- OMG.. text-to-how-to-video that is my dream... 1 click my personal how-to-video is created. It is all about context. Chatgpt is generative dialogue. But most online here in IH seem to be missing what this fact means to a "learner". Especially to all the kinesethic learners out there that most school systems have ignored. It means context is retained. I dont have to keep repeating myself. Learning is much faster when context is retained, and I get to dialogue. For example to chatgpt, you can say, nope your wrong, you have misunderstood my context -- for any question asked. LIke 200% more powerful than any search of database, book, pdf .. that is word match search.
AI models will boost both these trends.
catalystNewbie

·
4 years ago
·
Reply