Incredible Small 1.0 by Incredible - Agentic AI Models that can take 1000+ actions at once
byโข
RESEARCH PREVIEW โ Agentic AI Models that can take 1000+ actions at once, and work with large datasets. 100+ app connectors. Try the research preview assistant now at www.incredible.one. Built in Sweden, by Incredible ๐ธ๐ช
Replies
Andsend
Great work team Incredible
Incredible
@aage_reerslev1ย thank you Aage, happy you're enjoying it! :)
Congratulations on your successful launch.Wishing you great achievements ahead.
Incredible
@suchitra_biswalย Thank you so much! A very exiting start!๐๐ป
Congratulations with lots of wishes!
Incredible
@maira_syedย thank you a ton Maira! :)
Congrats on the launch, Philip and the Incredible team. Running hundreds of actions in parallel is impressive, what do you see as the main hurdle for teams to bring this into everyday workflows?
Incredible
@dhruv_patel59ย thank you Dhruv! We think there's 4 main things:
1. Repeatability/Reliability. (it doesn't do the same thing every time ๐)
Agents lack consistency in repeating the same objective. (2 different answers to same question = 2 different tax brackets selected for the same transaction. Funny, but unacceptable)
2. Cost. (Every AI-first scaleup's unit economics... is ... interesting. ๐ธ)
Data = cost. More steps = more money. More complex things = more money.
Agents are compute-intensive. Let's fix that.
3. Speed. (prevents every use-cases that has many steps ๐)
Agents are slow and limited due to token generation. Generating "JSON" schemas for taking actions is a pretty slow way of taking actions.
4. Accuracy (it just doesn't work sometimes ๐ฅ)
Agents fall short of human-level accuracy. Prevents them from just doing real work. There's a fix to this.
DiffSense
What's the top 3 coolest AI macros that you have seen people build with this?
Incredible
@sentry_coย hey Andrรฉ! there's a few insane quite simple ones. There was a self-learning blog agent, where it had access to Google Analytics to see performance of blog posts, a spreadsheet to do projections in, google trends + notion to publish them.
DiffSense
Oh! thats such a cool idea!!! You should put that on the front page! lead with use case, sell utopia. โจ
Incredible
so impressed by @philip_alm_at_incredible when we talked and on the platform they are building. bullish swedish founders rn (i.e. lovable, spotify, etc). can't wait to see their crazy ai growth chart...
Incredible
@pollockย thanks for the shoutout Tristan- super exciting times for sure!
Gilion
Very impressive. I really like the idea to build a product focusing specifically on making AI to take actions. This is an Incredible (!) start!
Incredible
@hlandgrenย Thank you a ton Henrik! Team is extremely excited looking at the use-cases that people are publishing!
DeepTagger
Love the vision ๐ค Is there a โsimulateโ mode to preview side effects before execution (dry runs, estimated API calls/costs)?
Incredible
@talshynย Thanks for the kind words!
Regarding "simulate" mode, not yet, but that is a great idea, and possible something we should explore in the coming weeks.
Incredible
@tommenย haha thank you Tommen. Happy you're enjoying it.
@philip_alm_at_incredible @emil_wagman_at_incredible @david_bacelj1 This is really interesting - just spent some time playing around and I think I understand the broad strokes / mass appeal for use cases.
I guess - as a theoretical user - where do you think someone should start - ie what processes should be first intertwined with the platform for folks to really get a grapple on the power of the platform.
I am finding - as a user of AI on a daily basis - it really helps (at least me) to have some smaller scale actions I could outsource (presume... writing my launch support comments with access to my PH profile / LinkedIn) - or is it more used to work alongside tools to parse down big data sets / user insights.
I quite like it - just curious how it might be applicable to me as a prospective user and where I might get the most value... again sorry if this info exists somewhere - I haven't dug in too deep.
Incredible
@dzaitzowย Hey Daniel, really appreciate you taking the time to dive in and for the fantastic, thoughtful question. You've hit on a key point โ figuring out the best entry point to really grasp what's possible.
To answer your question directly: It can absolutely handle both! You can definitely use it for smaller-scale actions, but its unique power becomes clear when a task involves multiple steps or connects different apps.
A great place to start is by automating a process that you do manually which involves moving or transforming information between tools. Think of a task that feels like "digital grunt work."
For example, a simple but powerful starting point could be:
"Take this meeting transcript from Google Docs, identify all the action items, and create a full project plan in Trello with assignees and due dates."
This starts with a single source of data (the doc) and expands into 100+ organized actions in another app. It's a perfect way to "get a grapple on the power" you mentioned because itโs a task that other AIs struggle with at that scale and reliability.
So while you could absolutely use it to draft launch comments, the real "aha!" moment might come from asking it to first analyze your top 10 most engaging LinkedIn posts, identify the core themes, and then draft 20 new launch comments tailored to those themes.
Hope this helps! Happy to brainstorm more specific use cases if you have a process in mind. Thanks again for the great feedback
@emil_wagman_at_incredibleย Super helpful and great to understand in a more granular capacity!