This is dead on arrival just like their, very good, powerpoint building product.
Nobody in the western world wants their value added closed source services (submit your H2 strategy deck to their cloud? Your contracts? I think not) especially when they are priced increasingly close to frontier.
Z.ai doesn't fully appreciate the level of distrust by their clients from western markets for simply being Chinese origin.
They do have credibility but it comes from being both open and having extremely capable models.
Removing one of these pillars makes a marginal risk / reward tradeoff probably a nonstarter.
Maybe LATAM, BRICS, Middle East are more willing to accept a closed / relatively expensive value prop. But not US / EU.
I mean forget the international intrigue angle. Some jackass troll with some money to burn can file a frivolous lawsuit against you seeking discovery of this material. Either the cost of production or the risk that they can find something embarrassing to pull out of context (or maybe even something debatable as wrong doing) means you'll pay some extortionary settlement to make them go away.
Beyond the obvious possible profit motive these attacks can be arms length arranged by competitors or people with an axe to grind to trip you up and slow you down. Because your activity is valuable to you there is an asymmetry-- they can bring the action cheaply because they aren't risking much, but your defense is expensive because if you flub it you could lose everything.
Part of the value of a legal advisor generally and particularly something like a legal AI is that it lets you map risk surface: "What if we started selling all the customer secrets? would we go to ultra-prison?"-- you don't ask because you're going to do it, you ask to perform a sensitivity analysis, to know where the dragons are hiding, and to setup contingencies. For stuff you're actually doing/going to do you want advice backed by a team who can defend it if it goes wrong.
All of hosted AI is a big risk from the perspective of discovery-- but particularly or legal advice/research that seems extremely radioactive and a bad tradeoff.
If anything being outside of the west might be a slight asset on this risk-- as there is some chance that they respond to a subpoena with "bite me"... but that probably doesn't overcome the other concerns you raise.
It can only be this. And it will fail the same reason Alexa failed.
Nobody will blindly trust Amazon to buy products when they know Amazon cannot and will not truly suggest the best product for the consumer.
Alexa was supposed to be the voice gateway to purchasing. But nobody trusts the result Alexa picks because it will not default to the best value to the user - only the best value to Amazon.
Recently it was estimated that Amazon makes $30-80 billion(!) per year off not showing you the economically best results by deliberately disallowing sorting on $ per unit [1] to force sellers into ads. Amongst other tactics (ex. filter by country of origin).
AI agents can sort on $ / unit to easily find you the best deal, rendering Amazon's various forms of paid placement irrelevant.
Extend this to filtering Chinese slop brands (also not able to be filtered) and the significant portion of revenue starts looking like knocking back margins to the 3-5% range. Basically an existential threat.
Amazon cannot allow AI agents and maintain current margins.
From what we hear / see in the community this will likely mean banning all abliterated / derestricted / decensored models on huggingface. And probably sooner than later.
Nvidia can't / won't tolerate headlines like "The scale of the problem is massive: Hugging Face, which hosts open source AI models, currently lists over 6,000 abliterated models."
In addition to this article, TC published another [1] just two weeks ago that agitates the same agenda.
Modelscope.cn seems to be one alternative (fewer reg-walls) but any others that have good selection of abliterated models? Especially those tuned to cyber / re team applications.
Alongside our MCP, skill, public API, .md files for all pages, llms.txt and https://fless.io/ai-agents info page we also wired up detailed analytics at every stage and for every request isolated by channel, both human and AI agent.
We track all agent calls at a pretty granular level to understand the agent funnel, by channel and provider (ex. Perplexity User, OpenAI User etc..).
We can't see what the agent did after calling us but we can see a funnel clearly which is useful to see where engagement and drop-off are.
Our first goal is for users to start a new apartment search - moving from 'landing page' - one of the surfaces mentioned above - to search created (ie. we're hunting for apartments on their behalf).
This has been helpful to understand agent behavior and where we need to improve skill / MCP etc..
Thanks for sharing!
In your case, since the skill terminates in an MCP/API call, you have a closed feedback loop.
I was wondering about skills that don't have that feedback loop, where the output the agent produced doesn't touch an external server. For instance, the frontend-design skill from Anthropic.
How would someone get insights into the user's experience of using that kind of skill?
We're working on an apartment hunting agent called Fless (https://fless.io).
Tell us what apartment you want.
We filter, sort and reach out to confirm apartments actually have a unit that meets your criteria by the move-in date you're looking for.
If they do, we schedule a tour for you and manage your calendar to ensure all the tours you book are nicely coordinated (with Uber rides scheduled if you need!).
No more emails, texts or spreadsheets and we also protect your privacy from post tour spam.
Once you find the unit you want we help you review the lease agreement, know what discounts to ask for and assist with scheduling movers who won't rip you off.
We're curious about how many people are using agents like Perplexity, OpenClaw, Hermes, Claude or others that support open skills and MCP to find an apartment.
We recently launched skills and MCP integration for agents at https://fless.io/ai-agents or via github / PyPI and other channels.
Check them out and let us know what you think - we respond to DMs!
Seems like this works in a very limited geographic area. Why not US-wide?
I did do this for a recent search. It scoured Craigslist listings, first evaluating must haves in text and looking at images for the promising listings. Crucially I wanted to first cast a wide net and fine tune my reqs as I learned what it turned up. The goal was quality over speed. I wish I could have used Zillow etc but they don't have a public API.
I'd pay for this and I'd recommend some kind of freemium approach. Interestingly I'd also like this for purchasing second hand stuff. I recently bought a TIG welder second hand. I had Claude compare various listings over the weeks to help judge price-quality. Would have been great to start a job and come back two weeks later with a top 3.
We've found the same. Agents are so sophisticated now that we haven't really focused on human users for our MCP and skill. Instead we have an agent page (https://fless.io/ai-agents) explaining where to find our MCP, what it does (finds apartments) and how to use it.
Combined with github presence and some other places agents just seem to find and use whatever works for them.
This is all largely invisible to end users who just use us through their agents to find apartments.
The exception is OpenAI which as you found out doesn't allow MCP for chat users (does for codex) instead requiring an app-store type listing. Not so open and triggers an app store style discovery process (and presumably tax in the near future).
We like the web + mobile app orientation (available without an app store!) along with the high degree of control.
Anyone moved between Paseo and T3? Curious to hear what motivated you in either direction.
reply