Thanks to all the kinda talented people who share their Hermes and other agent platform configs. I take the ones I find most interesting and feed them into my Hermes agent to assimilate!
Stop letting your AI agents post on X for you.
This place is for humans.
I'm all for using AI to help create content, but automating it 100% without even manually posting it is lame.
Keep X HUMAN.
(no bots allowed.)
🙏
This week marks a special milestone for Kong as we opened our newest office in Milan - a return to the city where our co-founders first started. 🍝
It's one of several milestones driving our momentum across EMEA:
🏢 New office in Paris + expanded offices in London & Budapest
📈 72% YoY growth in our EMEA customer base
👥 71% growth in our regional team over the past two years
You can learn more about Kong's growth across EMEA here: https://t.co/JeYPATetyu
I think about the AI context problem in three parts:
1) Raw enterprise assets (APIs, databases, docs, code) aren't usually in a format agents can use. The data exists, but agents can't navigate it effectively.
2) Identity and auth don't translate across the token path. A user authenticates once, but the agent hits walls the moment it crosses system boundaries with different SSOs or auth requirements.
3) Context doesn't evolve with use. Every agent interaction generates a signal, but it disappears instead of feeding back to make the context smarter over time.
The job is to take raw enterprise assets and turn them into context agents can explore and act on, while keeping governance intact.
You need to connect your sources, transforms them into agent-ready interfaces via MCP, handle identity translation across internal and external systems, and have a single registry to manage it all.
This pairs with your AI Gateway, which powers the API control tower to govern all AI traffic.
In the last few months at Kong, we've processed several trillion tokens through our AI gateway. Here's what we're seeing with enterprise AI usage:
1) AI request volume:
56% Anthropic
39% OpenAI
1.5% Gemini
(Huge surge from Anthropic over last 6 months to replace OpenAI as the leader)
2) Model preference is not uniform across industries.
For example, Claude is dominating in Finance, where the work requires complex, long-context workloads and high governance requirements.
3) The share of AI traffic through the platform has grown more than 6x in the last six months.
4) Average tokens/request is 6,621 and climbing higher as workloads deepen.
@NousResearch I already have a mature Hermes Agent installed on a headless workstation. Can I use this desktop app on my Mac to connect to the remote agent?
@higgsfield Getting CORS errors when trying to connect MCP:
Failed to load resource: net::ERR_FAILEDUnderstand this error
device?code=<redacted>:1 Access to fetch at 'https://t.co/s4A5No0vak' from origin 'https://t.co/5fW3xOQJiH' has been blocked by CORS policy
We’ve spent decades perfecting APIs, event streams, and integrations… and yet one challenge keeps showing up: context.
Without it, even the most advanced systems behave like confident guessers. With it, they become truly intelligent. 🧵👇
@WesRoth How do I get access to seedance 2 please? Is it still closed beta?
Also huge fan thanks for all the work you put into your material here and on YouTube
@kong wouldn't exist without @travisk and for that we are very grateful during those hard and heavy days at the beginning of our journey.
Another part of the story is how Travis helped us find our first real office at the DogPatch in 2010 (co-working space at the Pier 38 in SF where @instagram also went GA one of those late nights at the office) by emailing and asking to the office manager to help these two Italians who had just raised some money so "they can upgrade from plain Ramen to flavored Ramen".
All the help that Travis gave us came with no string attached, he genuinely wanted to help founders from all walks of life, and I will forever be grateful to him.