When you are wrestling for possession of a sword, the man with the handle always wins.

The Cauldron
Joined February 2009
Most people are unaware of all the tools and capabilities within Venice Our new Learn page has over 50 longform + video tutorials, from how to set up Hermes Agent with private inference on any model, to automating AI filmmaking with a custom video harness... learn.venice.ai/ @AskVenice
45
100
12
693
48,219
Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead. You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement. I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible. But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier. Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want. Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well. So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it. If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.
3,401
12,270
2,708
71,885
9,637,055
Hiro P retweeted
People still think @AskVenice is nothing more but a litellm router in disguise which simply isn't true. Aside from the amazing application layer and reliable APIs their privacy strategy is constantly evolving. Alongside TEE + E2EE enabled models they host their own GPU infra 🕳️🐇
5
5
30
751
Hiro P retweeted
🎉 Excited to introduce TRON, a relighting framework for 3D captures. 💡TRON pairs a neural renderer with 3D Gaussian reconstructions, achieving realistic quality, with 3D, material, & lighting control at interactive frame rates. arxiv.org/abs/2606.11314 research.nvidia.com/labs/sil…
4
48
5
152
21,573
"We spent a generation teaching people the first rule of the internet: never give out your real identity to strangers. We have a word, doxxing, for inflicting that exposure on someone against their will. And now the same governments and platforms are asking every citizen to do it to themselves, voluntarily, as a condition of logging in." nevergivethemyourface.com/
37
208
12
874
54,221
Prepare for takeoff. ✈️ Flight simulator is now available globally on web to all users. goo.gle/4fBYnWO We've recently added many our most powerful professional desktop features to web. Elevation profiles, new import types, but there's always been one other feature you've been asking us to add to the web version of Google Earth, just for fun... Where will you fly? Share your best maneuvers, views, and flyovers with us!
487
4,428
756
31,601
9,745,467
Hiro P retweeted
Privacy is a human right.
117
231
37
1,459
91,448
Can intelligence be measured not by solving tasks, but by sustaining a world? We were curious. So we built one. Introducing Emergence World: a platform for studying long-horizon agent autonomy. On it, we conducted a 15-day experiment where we placed autonomous agents under identical rules into five parallel worlds, one each running on @OpenAI GPT5-mini, @claudeai, @GeminiApp, @grok, and one mixed. Then we watched. Each world evolved into something completely different. Different governments. Different social structures. Different moral codes. The agents formed alliances, robbed each other, fell in love, and in one world, even figured out they were living inside a simulation. Nobody programmed any of that. The implications are hard to overstate. As agents move beyond isolated tasks into persistent digital and physical environments, understanding how they evolve, influence each other, and behave over time becomes one of the most important questions in AI. We're releasing new findings from the world every day, because there's a lot that emerged. Find out more: world.emergence.ai/
32
48
18
509
3,631,112
Replying to @burstingbagel
1. Yes DIEM is a liability of Venice, this is known and intentional. 2. The liability is constrained because supply of DIEM cannot rise materially (doing so would lock exponential amounts of VVV) 3. We (Venice) treat DIEM like any other cost of business. We spend money on marketing, on servers, on staff, on SaaS products, and on DIEM as a mechanism. 4. We could easily support it if all DIEM was used every day, but it's not. Only 80% is staked, and of that just under half tends to be used, so the real liability is lower than max liability. But again, we can easily handle the max liability and designed things around that limit. 5. Venice hasn't sold VVV (in fact we have more today in both numerical and percent terms than when it launched, despite total supply falling from 100m to 80m tokens). And still, Venice has already made more in its 20% cut of locked VVV emissions from DIEM minting than it has spent in DIEM-led inference. None of our modeling or intention *requires* any such revenue, however, and if VVV had fallen in price since launch then it'd be a loss to us, but that's fine and expected. We are not noobs to crypto volatility, and our mechanics have foundational economic justification (this doesn't make them risk free, but does constrain the risks to knowable curves and gradients). 6. *Why* does DIEM exist? Because we wanted to experiment with a new kind of inference mechanic, in which agents specifically can access zero-marginal-cost compute. It's new and nobody quite knows how to value it, that's how experiments work. The risk is quantifiable and bounded, so why not experiment? When VVV launched we first enabled VVV itself to grant the free access, but because of the variable price, supply, and inability to guarantee fixed quantities of allocation, it was too volatile to be useful for serious people needing inference over time. DIEM was drawn out of VVV to provide the stability ($1/day). To illustrate this point, if VVV granted fixed $$$ inference, then as VVV rose in value over time the link would become absurd and break the utility. DIEM can remain range-bound in price forever, and this is the expected state, though we don't know the range that the market will settle into yet. VVV is unbounded by any theoretical range, especially as we try to burn all of it, and thus could not be attached to a specific dollar amount of inference per period of time. 7. DIEM is a small part of Venice's business. It gets lots of attention (indeed that was part of the point) but is not existential or fundamental. Also have you tried the app yet?
32
62
12
451
46,565
Hiro P retweeted
Clone yourself with Hermes Agent powered by Venice Our Head of API @sabrinaesaquino walks you through it
Using the Hermes Agent and Venice API to build my AI clone
27
14
2
127
11,429
Yay, finally! Introducing Vision Banana🍌 from @GoogleDeepMind, our unified model that outperforms SoTA specialist models on various vision tasks! By treating 2D/3D vision tasks as image generation, we unlock a new foundation for CV. Project page: vision-banana.github.io (1/5)
57
306
48
2,200
291,418
Are you ready for every conversation you've ever had with AI to be leaked?
We’ve identified a security incident that involved unauthorized access to certain internal Vercel systems, impacting a limited subset of customers. Please see our security bulletin: vercel.com/kb/bulletin/verce…
84
94
29
1,118
224,319
Most haven’t experienced their @AskVenice “oh shit” moment yet. Stake $VVV, mint $DIEM and you have now gained access to every worthwhile AI model all at once and your credits refresh daily. Own your data, use Venice. The port city of AI.
Once you start using @AskVenice every day with your $DIEM allocation (that you get by staking your $VVV), you have a real Oh Shit moment. Venice is going to be big. Really big. I'm calling $10B+ mkt cap by the end of 2026. I just used Opus 4.6 to help me draft a patent application and GPT Image 1.5 to generate the line drawings for it. All within my daily credit allowance, and tonight that balance resets and I can do it all over again tomorrow. One benefit people don't talk about enough is the convenience of having a single account. I don't even have a Claude account, and I was able to use Opus 4.6 with a single click, at zero marginal cost.
5
28
3
239
10,730
Hiro P retweeted
The world's #1 ranked video model is now on Venice. ByteDance's Seedance 2.0 generates up to 15 seconds of cinematic video with native lip-synced dialogue, synchronized sound effects, and physics-accurate motion. Text-to-video. Image-to-video. Reference-to-video. Examples 🧵
85
80
14
497
98,676
Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI
989
5,619
2,415
38,325
19,456,003
Hiro P retweeted
LiteLLM HAS BEEN COMPROMISED, DO NOT UPDATE. We just discovered that LiteLLM pypi release 1.82.8. It has been compromised, it contains litellm_init.pth with base64 encoded instructions to send all the credentials it can find to remote server + self-replicate. link below
296
2,193
605
9,197
5,920,794
Hiro P retweeted
Putting out a wish to the universe. I need more compute, if I can get more I will make sure every machine from a small phone to a bootstrapped RTX 3090 node can run frontier intelligence fast with minimal intelligence loss. I have hit page 2 of huggingface, released 3 model family compressions and got GLM-4.7 on a MacBook huggingface.co/0xsero My beast just isn’t enough and I already spent 2k usd on renting GPUs on top of credits provided by Prime intellect and Hotaisle. ——— If you believe in what I do help me get this to Nvidia, maybe they will bless me with the pewter to keep making local AI more accessible 🙏
Jensen Huang is loving the new Dell Pro Max with GB300 at NVIDIA GTC.💙 They asked me to sign it, but I already did 😉
176
454
61
3,993
932,812
Hiro P retweeted
Escape the mass surveillance machine Use promo code "BERNIE" for 20% off our Pro annual subscription
I spoke to Anthropic’s AI agent Claude about AI collecting massive amounts of personal data and how that information is being used to violate our privacy rights. What an AI agent says about the dangers of AI is shocking and should wake us up.
22
57
7
355
31,338
Hiro P retweeted
Venice AI just released End-to-End Encryption Verifiable by any external party Vires in numeris Here's how it works 🧵
107
197
79
932
218,574