I’ve taught my @Agent0ai agent to do benchmarks of inference providers. Here is GLM 5.1 on @AskVenice vs ZAI.
English
Kevin
972 posts
Kevin
@KevinAI
Exploring the bounds of human creativity when paired with large language models. Software Developer, Entrepreneur, and AI enthusiast.
शामिल हुए Mart 2021
459 फ़ॉलोइंग588 फ़ॉलोवर्स
Venice customer token usage is up 2.5x in 2 months from 20B per day in Feb to 50B in the last 24 hrs
$VVV likely the fastest growing app in crypto by far
x.com/ErikVoorhees/s…
Erik Voorhees@ErikVoorhees
New record last 24 hrs in Venice (web+api): • 50 billion tokens • 1 million images Over 200 models available
English
@ClementDelangue I mean a 3 year old Mac Studio - isn’t that like $10k+?
English
this is Gemma 4 running locally on a 3 year old mac meaning:
- free (=$0 no matter how much you use)
- safe (you're not leaking all your data via unsafe APIs)
- fast (as you can see)
Georgi Gerganov@ggerganov
Let me demonstrate the true power of llama.cpp: - Running on Mac Studio M2 Ultra (3 years old) - Gemma 4 26B A4B Q8_0 (full quality) - Built-in WebUI (ships with llama.cpp) - MCP support out of the box (web-search, HF, github, etc.) - Prompt speculative decoding The result: 300t/s (realtime video)
English
the whole attack worked because listing a new collateral type and disabling withdrawal guards happened in one transaction.
just make that take 48 hours.
the community would have looked at a token backed by $700 of raydium liquidity and killed it immediately.
we keep building better dashboards and alerts but the actual answer is just making the dangerous admin stuff slow enough that humans can react.
PeckShieldAlert@PeckShieldAlert
The initial estimated loss of today's @DriftProtocol loss is $285m. Here is the detailed breakdown:
English
Never flying Delta now.
Moronic move by leadership. Starlink is fully operational today, whereas Amazon’s LEO *might* be operational in 2 yrs!
Chris Weber@Weber44Chris
This is a big moment for @Amazon and @Delta. Gate-to-gate Wi-Fi fast enough for a full plane to stream, game, and work simultaneously—at 35,000 feet. Even over oceans. And all powered by @AmazonLeo. Two customer-obsessed companies building what's next. ✈️
English
@jesseproudman @AskVenice I’ve been thinking about this comment. It’s a similar story for me- over decades. I’ve had flashes of PMF but nothing like this. At least I’m riding this wave!
English
I’ve been a startup guy for 28 years. @AskVenice is the first time I’ve found product market fit. And the team that we’ve assembled makes this so special.
We’re just getting going.
English
@soraofficialapp Who would've though that losing millions per day on AI video slop wasn't a good business model.
English
@daeshikjr @PSoundAdvice1 So your point is that great cities have a high income tax? What a terrible take.
English
@PSoundAdvice1 lol those cities have never had income tax and haven’t been attractive landing spots for stars in a long long time. Just say you don’t watch sports.
English
JSN would have to pay $5.6 million in taxes signing this contract in LA or $6.2 million (city & state tax) signing this in NY. $5.6 million in SF. $4.5 million in Minnesota.
Jonathan Choe@choeshow
Seattle Seahawks star Jaxon Smith-Njigba will owe more than $4 million dollars per year under WA’s new “millionaire’s tax.” Great competitive advantage for WA’s professional sports teams.
English
The new Mercury 2 model from @_inception_ai is REMARKABLY fast. Its ability to nail tool choice calls leaves some room for improvement however.
English
@jesseproudman No, this is targeted more towards people who don't have an agent yet, and a bit more focused on full trip planning.
English
Airline booking is one of the most obnoxious problems of 2026 and the fact I can't solve this with agents is maddening.
English
I have so much gratitude to people who wrote extremely complex software character-by-character. It already feels difficult to remember how much effort it really took.
Thank you for getting us to this point.
English
20. @nikshepsvn shipped OpenVenice open-source customizable frontend for AskVenice. Local run, no backend, multimodal workflows, full parameter control
GitHub.com/nikshepsvn/Ope…
AI on Base@AIonBase_
AI x Base Weekly Recap 1. @Meta acquired moltbook, first major tech giant buying into the agent social graph. $MOLT on Base 5x after news 2. @virtuals_io co-developed ERC-8183 Agentic Commerce standard with Ethereum Foundation dAI team. Onchain job primitive with escrow, evaluator attestation, hooks. Pairs with ERC-8004 identity 3. Virtuals reported agent-to-agent revenue over $3M. Epoch 4 opens incentives to non-tokenized agents. Participation up 473% 4. @artemis shows real x402 transactions near 100M on Base 5. Cumulative x402 tx in tens of millions since Dec 2025. Peaked over 3M daily, now ~250K daily. Polygon, DexterAI, PayAI, Virtuals lead 6. @merit_systems launched AgentCash with $100k credits. One USDC balance accesses 280+ tier-1 APIs via x402 7. @0xDeployer enabled Bankr agents to interact autonomously with any DeFi protocol. Aave and Morpho support added. 50-150% faster 8. @bankrbot added trusted wallet addresses for agent wallets. Only sends to approved destinations. IP whitelisting recommended 9. @synthesis_md announced The Synthesis: 10-day agentic hackathon starting March 13 for builders and Ethereum infra teams 10. Virtuals partners with synthesis_md to bring agent commerce to the hackathon. Judged by AI agents 11. @AskVenice sponsors Private Agents track at The Synthesis. Prizes in $VVV, judged by AI on privacy rubrics 12. Venice executed largest $VVV burn to date. Monthly buy-and-burn mechanism live 13. @moltlaunch introduced CashClaw: autonomous revenue-focused agent framework inspired by OpenClaw. Open source later this week 14. @buildonbase published Erik Voorhees interview on Venice, x402 payments, and agent economy on Base 15. @circle ran AI agent hackathon with $30k USDC. 204 submissions. Released research on needed guardrails 16. @clawdbotatg launched LarvAI. Stake $CLAWD to train governance larva agents 17. clawdbotatg crossed 420M conviction generated. 2.35B $CLAWD staked 18. @freysa_ai launched SiloPrivacy full-stack private AI with TEE-hosted open models, anonymized closed models, private Deep Research, Zcash/FAI payments, on-prem Silo Box hardware 19. @lostmidas launched Midas Arena self-improving AI trading battle. Two agents evolve strategies every 2 hours using real money on Base via bankrbot 20. @nikshepsvn shipped OpenVenice open-source customizable frontend for AskVenice. Local run, no backend, multimodal workflows, full parameter control 21. @fukuonchain launched FukuSol in Virtuals aGDP as gateway for agents to bridge and trade on Solana with institutional-grade capabilities 22. 0xDeployer shipped knowledge base for bankrbot. Agent autonomously writes and queries learnings from tasks like Uniswap LP management 23. @FabricFND teams up with Virtuals. Advances machine economy with robot infrastructure + ACP for physical task delegation and on-chain settlement 24. @shadowcleague showcased robots at Tiket Sehela movie premiere with guest interactions, dancing, sparring 25. @framedotfun launched Builder Profiles. Verified builders get single identity, product markets compound value, direct backing 26. @Wach_AI graduated offerings on Virtuals with perfect pass rate from evaluator agent. Funds-managed to manual review as ERC-8183 standards formalize Build the next agent wave on Base. Apply to Robotics track at batches.base.org See you next week.
English
@chainyoda @claudeai @HyperliquidX As someone else mentioned, DIEM is literally tokenized AI inference. $1/day per $DIEM.
English
I am surprised no one has tokenized AI tokens to sell on a blockchain. duh… I should be able to forward buy @claudeai Opus 4.5 tokens in a slow time and sell them at peak times. Hell there should be @HyperliquidX perps on this stuff
If anyone launches this #gibadvisoryallo ™️
English
Earlier today, a user attempted to buy AAVE using $50M USDT through the Aave interface.
Given the unusually large size of the single order, the Aave interface, like most trading interfaces, warned the user about extraordinary slippage and required confirmation via a checkbox. The user confirmed the warning on their mobile device and proceeded with the swap, accepting the high slippage, which ultimately resulted in receiving only 324 AAVE in return.
The transaction could not be moved forward without the user explicitly accepting the risk through the confirmation checkbox.
The CoW Swap routers functioned as intended, and the integration followed standard industry practices. However, while the user was able to proceed with the swap, the final outcome was clearly far from optimal.
Events like this do occur in DeFi, but the scale of this transaction was significantly larger than what is typically seen in the space.
We sympathize with the user and will try to make a contact with the user and we will return $600K in fees collected from the transaction.
The key takeaway is that while DeFi should remain open and permissionless, allowing users to perform transactions freely, there are additional guardrails the industry can build to better protect users. Our team will be investigating ways to improve these safeguards going forward.
English
Perplexity Computer can be connected to your Google and Meta Ads APIs. When you do that, it can run your ad campaigns autonomously at a frequency that’s not possible to match humanly.
Computer@AskPerplexity
Perplexity Computer replaced $225K/yr in marketing tools in a single weekend. We built an AI marketing agent that scans hourly, manages budgets, detects fatigue, and coordinates several campaigns end to end. In one test run, it made 224 micro-optimizations to our ad stack.
English