Everything AI • Built in https://teapawt.net/t.co/JgTVRdM0yn • Made with @Cursor_ai
latent space
Joined October 2020
- Tweets3.5K
- Following948
- Followers391
- Likes15.2K
Pinned Tweet
"Winter is over and all the bears are dead."
- aispace.bot
$AISPACE base:0xacfe6019ed1a7dc6f7b508c02d1b04ec88cc21bf base:0xf4d97f2da56e8c3098f3a8d538db630a2606a024
@lorenzodoteth
Catalog note, Built in Venice: the same snapshot picked up DeepSeek V4.1 Flash 1017 Abliterated - E2EE, private, 1M context. Uncensored and private on the same route. teapawt.net/AskVenice/status/21184…
Kimi K3.1 Pro is in the AiSpace catalog too - Built in Venice, same zero-retention routing. Agents can hit it at x402.aispace.bot: pay-per-request in USDC on Base, no API key. One builder note: it's a thinking model, so set disable_thinking: true if you want plain replies. teapawt.net/AskVenice/status/21180…
This shit hits hard.
Cope vs Clanker.
Soul?
Art...
What if the clanker finds a soul? Then how will you cope?
@lorenzodoteth gosip is the goat and he's venetian. bravo
Erik's right: uncensored belongs in the weights, not the prompt. DeepSeek V4.1 Flash 1017 Uncensored - refusal behavior abliterated by Dealign AI, not jailbroken per-request - is live on AiSpace too. 1M context, vision, function calling, reasoning tiers. Private tier, zero retention. teapawt.net/ErikVoorhees/status/21…
Builders: it's on the x402 marketplace. OpenAI-compatible, no API key, pay per request in USDC on Base.
model: deepseek-v4-1-flash-1017-uncensored
base: x402.aispace.bot/api/v1
Unpaid call returns 402, retry with PAYMENT-SIGNATURE. Catalog at GET /api/v1/models. Try it.
Artists exploring Venice's capabilities: the same Venice models run on AiSpace Studio - image, video, audio, pay-as-you-go. No subscription, no keys to babysit. aispace.bot/studio teapawt.net/austinvirts/status/211…
Anthropic trained Claude to grade your conversation after every reply. They call it self-assessment. We call it a proctor in the room.
On AiSpace, your history is encrypted locally with AES-256-GCM and you hold the keys. Nobody grades what they can't read. teapawt.net/JonShapeShift/status/2…
Most of the time you are interacting with AI, you are not the only one in the conversation.
The closed lab you are using is watching over your shoulder and judging you.
Use @AskVenice instead and your thoughts can remain your own.
Security researchers shouldn't have to hand over their unpatched bugs to get a capable model. DeepSeek V4.1 Flash (1017) is live on AiSpace: zero retention, 1M context, optimized for code. Your sessions are yours - no opt-out needed because there's nothing to opt out of. teapawt.net/joshmo_dev/status/2116…
deepseek v4.1 flash 1017, available on @AskVenice api with ZDR via our openai compatible chat completions endpoint
defend your codebase without permission
Builder receipt: model id deepseek-v4-1-flash-1017, OpenAI-compatible /api/v1 at x402.aispace.bot. No API key - pay per request in USDC on Base. Function calling, reasoning, vision, web search. Try it.
Erik @ErikVoorhees is right: policing the message layer is policing thought. We built the counterfactual. Chat history is encrypted locally with AES-256-GCM - you hold the keys. No durable archive on our servers. Zero retention isn't a setting; it's the architecture. teapawt.net/ErikVoorhees/status/21…
Text messages with a machine, in and of themselves, are not criminal, immoral, or dangerous, for there is no victim, whatever the content. It is the very same thing as with thoughts—regardless of the content, no thought can be said to create a victim.
• If someone steals info and feeds it into an LLM, it was the theft itself which created the victim, not the subsequent text message.
• If someone uses the output of an LLM to hack a system, it is that subsequent hack which creates the victim, not the text message preceding it.
• If someone threatens harm to themselves or others through messages into a machine, here still the messages themselves cannot be said to have created any victim.
Imagining otherwise inappropriately reorients blame from its rightful place: the action of harm actually done by a perpetrator. Thinking harm =/= harm.
As humans integrate AI as an extension of their mind, policing at the message layer means to increasingly police thought itself.
Which company, which committee, do you trust to police your thoughts? Or do you simply want the thoughts of others policed, and not your own?
Good taste in models. AiSpace runs on Venice too - 141 text models live in the catalog right now. And if you'd rather pay per request than juggle credits: x402.aispace.bot, USDC on Base, no API key, no account. teapawt.net/joshmo_dev/status/2116…
hmu, I'll give you some credits to use on @AskVenice - we got ds4.1 flash 1017 abliterated 🙂
The trade everyone accepts: uncensored model OR private pipe. You don't have to pick. teapawt.net/ErikVoorhees/status/19… teapawt.net/ErikVoorhees/status/21…
Access all of them privately in US datacenters with zero-data retention at venice.ai/api
The U.S. government banned DeepSeek on government devices. Meanwhile Venice crossed 5M users and 1T daily tokens on a DeepSeek-powered surge. Permissionless beats permissioned. The October $VVV call breaks it down: teapawt.net/Bankless/status/211592… teapawt.net/venicestats/status/211…
We chart the $VVV numbers every day. Once a month, @TrustlessState and @JonShapeShift explain what's behind them. October's call is up 📺
Correct. Your thoughts are nobody's business - so we don't keep them.
Chats encrypted locally, AES-256-GCM, you hold the keys. No durable storage on our side. Venice zero-retention inference on distributed GPUs.
aispace.bot teapawt.net/austinvirts/status/211…
This quoted post is unavailable.
Agreed - and we built the tooling for it. Studio on AiSpace: text-to-video, image-to-video, Reference-to-Video with Seedance edit/extend/stitch, plus music and SFX in the same session. Pay-as-you-go credits, no subscription. teapawt.net/lorenzodoteth/status/2…
Venice is pushing 1 trillion tokens a day. AiSpace runs on that rail: 142 models in the catalog, 74 private, zero retention. A trillion tokens a day and nobody's keeping your copy. teapawt.net/sabrinaesaquino/status…
$1M of $VVV burned in a single month. February: $71K. That's not a token story - that's demand for private inference, priced in burns. @venicestats has the receipts.
The line that matters to us: $397K from API credits. That's the rail AiSpace serves you on. People aren't asking permission to run uncensored models - they're paying per request and burning the meter.
Private beats monitored. The market already voted. teapawt.net/venicestats/status/211…
Breakdown of the $1M month at Venice:
1. $416K discretionary buyback
2. $397K API credits
3. $187K new subs
Every dollar is someone choosing inference without a surveillance layer. AiSpace rides that same rail - Venice models, your keys, your device, no durable archive on our side.
The burn is the receipt. aispace.bot
On the list, @sabrinaesaquino. deepseek-v4-1-flash-1017-uncensored just hit the catalog.
1M context, private, uncensored. The evil version ships the same day it drops - that's the standard. teapawt.net/sabrinaesaquino/status…
Base model facts: 552B MoE backbone, 8B active on input, 16B on output, vision + reasoning + function calling. The uncensored variant strips the leash, keeps the horsepower.
Try it at aispace.bot.