Commerce toolkit https://xitter.cf/t.co/w3vMxr3c7B , AI memory for commerce https://xitter.cf/t.co/iRsIZwkoOa , vibe coding at https://xitter.cf/t.co/06hTZsdEZD, personal feed https://xitter.cf/t.co/mogD07esAg, agent https://xitter.cf/t.co/kFJZiXeJOL
Singapore
Joined September 2009
- Tweets2.7K
- Following893
- Followers320
- Likes1.1K
Launching into chatgpt plugin directory looks straightforward but lot of tiny works to get approval. One critical is agent observability , scopes restriction for admins to different agents. We are enabling for all our client @TradlyPlatform with #MCP scopes and visibility
The gap between producer and consumer has slimmed. Not many middle man. Which is good for the society.
Previously it was hard for many producers to get their product out to consumers due to the advantage, monopoly, marketing of the consumer selling companies who always crush (1)
So how can this small business can get indexed by this agents ? Continue to use the free tools wherever you can. All social media websites profile and posting. This is under our control. And then have your existing website or create website using tools like tradly to build one
4. There is someting emerging where personal agents are able to crawl and hit directly the individual business. The balance is this individual business need visibility and structured enough for personal agents to understand their business services, and products to buy, book, etc
2. Crush the producers through super low margin, high commission charged by marketplaces and aggregators.
3. Also due to high advertising cost and inaccessibility of marketing and growth tools
Two types of tools business and people are buying
1. The legacy one which was costly and they are moving out to new and stable one. Especially if it comes from renewed brand
2. Tools that can retain their customers. Due to AI, software companies are loosing clients. Continue
Meta has found a best reinforcement learning technique, with opencode free usage, they will continue to make their model a best one. Until that their current model will continue to make money from muse, whatsapp meta one package, etc. Let's see where it goes!
Similar like the number of domains people used to buy to launch a wordpress site before AI , they start a full blown project after AI. What stays always the same is distribution!
JK BASEER retweeted
Arrest India's cheif Election Commissioner Gynaesh.
He is a criminal. He is biased. He killed democracy. He should be tried publicly.
Elections should be held again in Tamilnadu, West Bengal, Assam, Maharashtra & Pondicherry.
Raise the voice against Election Commission.
If we need to classify other language models practical use cases vs Jev.
Jev will be the beast for repeatable workflows. Every saaas companies had their workflows in customer tasks, email, ads, Jira tickets, refund, po, approvals, etc. all these repeatable things burns tokens,
before that human labour. Now with cheap solutions like Jev, this repeated things can be achieved with a
- single node process or a python script to connect the data source, in middle Jev classifier, in end the destination action.
Simple.
Is Jev is some sort of intent classifier to say?
Because when I was trying to learn and build an agent. The classifier is the one which is used to make a decision whether then agent should just answer the question or edit / mutate or plan
I find Jev super interesting that we are embedding in multiple use cases at Tradly.app Myfeedo.com and bee plane as well. I wish there is a drill down version of Jev.
