LATEST: ⚡ Nigeria accounts for roughly 60% of stablecoin inflows in sub-Saharan Africa and received about $59B in crypto inflows from July 2023 to June 2024, according to the IMF.
Looks like the options thing is happening already!
See also: various people thinking through and building different versions of the idea in the thread: https://t.co/gFNEvCbHct
Though I do strongly urge that if any of these get on mainnet quickly, we formally verify it first. I hope @vyperlang and/or https://t.co/OMFlWRqJda folks ( @Fricoben) can help!
(Also, now is a good time to be thinking about robustness-optimized oracles)
https://t.co/j1dxLV4Pn4
Crypto has been walled off from the real economy for years. That's changing. @eddylazzarin joins @smc90 and @rhackett to break down why crypto is entering a completely different phase, and what gets built once the rules finally catch up to the technology.
00:00 Intro
01:25 What it means to be a GP
02:00 Consensus vs. non-consensus bets
04:27 Network tokens and the CLARITY Act
09:35 Revenue, value capture, and network-token business models
21:18 Stablecoins as crypto’s first killer app
28:52 Engineer-philosopher mindset
39:04 Intellectual influences
52:40 Eddy’s path to crypto
01:03:15 The exuberant adoption phase of AI
01:14:38 Being "pro–AI psychosis"
Messari has been acquired by Blockworks.
This brings together the two largest crypto data and market intelligence platforms to build the trusted information infrastructure onchain markets need as they mature.
The biggest marketplace on earth won't be assets humans own. It'll be the data AI needs. And we won't see any of it.
I sat down with @jvisserlabs to map out Reed's Law, the invisible economy, the AI supercycle, and what it all means for capital markets. As ever, please enjoy!
Good take
My guess is
- demand for intelligence is near infinite
- but 80% of workloads will be running on 99% cheaper models within 12-18 months
- 20% of workloads will still run on latest gen models where IQ maxing is important (scientific breakthroughs, higher level ochestrator agents?)
- rough analogy might be what % of macbooks or gaming PCs sold have the maxed out specs for CPU/GPU, prices are falling much faster than Moore's law here though
- this leads me to think the limiting factor will be energy and compute, not better models
At Coinbase we're working hard on routing prompts to cheaper models where appropriate, and in some cases have been able to keep costs roughly flat, while token usage continues to grow exponentially.
Join the AGI community for the 19th Annual AGI Conference (AGI-26) in San Francisco, California, from July 27 to July 30, 2026.
This premier gathering brings together leading researchers, industry pioneers, policymakers, and innovators from around the world to discuss cutting-edge research, collaborate, and shape the future of intelligent machines.
Experience keynotes from visionary leaders, including Ben Goertzel, Neil Gershenfeld, Alison Gopnik, Karl Friston, Gary Marcus, Michael Levin, Alexander Lerchner, with more speakers to be announced.
Register now to secure your ticket or watch online: https://t.co/v6EGo5m7rC
Excited to share that I’ve joined OpenAI in London to work on pretraining!
I’ve spent the last few years on pretraining and long-context, and I’ll be helping grow the London team. We’re hiring exceptional researchers, please reach out: [email protected]
Before the week ends, let's acknowledge one of the most INSANE week ever for open AI, with 25+ notable open-weight drops across every modality:
🧠 LLMs
→ NVIDIA Nemotron 3 Ultra: 550B hybrid Mamba-MoE, only 55B active, 1M context, MMLU 89.1. NVFP4 variant claims ~5x throughput on Blackwell. First openly-weighted 550B hybrid Mamba-Transformer, closing the gap with frontier closed models.
→ Google Gemma 4 12B: fully open dense any-to-any (text/image/audio/video), 256k context, encoder-free, 140+ languages, AIME 2026 at 77.5. Shipped with a 23-checkpoint QAT wave (mobile ONNX + MLX). Most deployable model of the week.
→ StepFun Step-3.7-Flash: 198B sparse MoE VLM, ~11B active, SWE-Bench PRO 56.3. Apache 2.0.
→ Liquid AI LFM2.5-8B-A1B: edge MoE, just 1.5B active, 128k ctx, MATH500 88.8, MLX-ready. Best on-device option this week.
→ JetBrains Mellum2-12B-A2.5B-Thinking: their first open MoE, near-Qwen3-14B coding at 2.5B active. Apache 2.0.
🎨 Image gen (the surprise of the week)
→ Ideogram 4: their FIRST-EVER open weights. 9.3B flow-matching DiT trained from scratch. #2 overall behind GPT Image 2, top open-weight model on Design Arena + LMArena. Strongest open checkpoint for text-rich images, full stop. It has taste. Still can't believe this is open weights.
🔊 Audio & Speech (a breakout week for open TTS, 4 labs shipped)
→ Boson Higgs Audio v3 4B: 102 languages, 21 emotions, singing/whispering/shouting, sub-second TTFA.
→ RedNote dots.tts: the only fully continuous (no codec) open TTS pipeline, Apache 2.0.
→ Google Magenta RealTime 2: real-time music gen, <200ms latency, text+audio+MIDI. multimodalart ported it to PyTorch within hours with live ZeroGPU demos.
→ NVIDIA Nemotron-3.5 ASR: 600M streaming, 17x more concurrent streams vs Parakeet RNNT 1.1B.
👁️ Vision & VLMs
→ PaddleOCR-VL-1.6: SOTA document parsing at 1B params, Apache 2.0.
→ Baidu NAVA: 6.3B joint audio-video gen, best-in-class A/V sync, Apache 2.0.
🎬 Video, 3D & World Models
→ NVIDIA Cosmos3-Super: 64B omnimodal world model coupling action trajectories with video+audio gen, for Physical AI.
→ JD JoyAI-Echo: up to 5-min multi-shot text-to-video on LTX-2.3.
→ ByteDance Bernini-R + VAST TripoSplat (single-image-to-3D Gaussian splats, MIT).
I spent the entire day researching the Zcash issue.
I spoke to devs, to @mert , to one of the founders @EliBenSasson .
My conclusion is the protocol is *probably* safe and we got away unscathed and that Zcash has a very bright future..
Here is why:
1. There is very little chance that the vulnerability was exploited, why? You would have needed to know the protocol really well and you would have needed sophisticated AI tools that were not available.
https://t.co/TSmHPjuMkw the same time the behavior or the shielded pool and the price are commensurate with the assumption that it was not exploited. There was no selling out suspicious outflows.
3. We will only know with a higher probability when the new shielded pool is launched and we don’t have “over-migration” issues.
4. If we survive that then with the new upgrade that make Zcash quantum resistant, you will be able to verify cryptographically that there are no future issues in the shielded pool and from that point this can’t happen again. (Or at least we can pick it up in the shielded pool”
When this happens Zcash is a privacy protocol, with a Zk shielded pool , that is as safe a a regular blockchain.
In a nutshell , there is one more hurdle ahead and that is the migration to the new shielded pool and if that happens without issue, we are in the clear.
The Rally That Wasn't
$BTC has fallen to $66K, with profitability collapsing, realized losses surging, and spot sellers regaining control. We are below the ETF cost basis, and options continue pricing elevated risk.
Read the full Week On-Chain👇
https://t.co/IQnTQMH0jr
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
INSIGHT: $ADA falls below $0.20 for the first time in five years as @IOHK_Charles says he is "taking a break" and warns of a coming "wave of failures" in Cardano's ecosystem after the community cancelled its flagship 2026 Summit and rejected key treasury proposals.