BTC is still within reach of the open CME gap near $81,145, but I am not treating this as a clean all-clear. Price has managed three days above overhead resistance and the April rally has been real, yet the chart is starting to look like a pivot high setup right at the top of the cloud. If BTC gets one more squeeze higher and then rolls over, that would fit the kind of max-pain move I have been warning about.
ETH looks much weaker than BTC here, and that is the bigger warning for the rest of the market. It failed again at resistance, the support fan is already broken, and I still think a move back toward $2,000 makes more sense than a fresh breakout. BTC.D keeps pushing higher, stablecoin dominance still looks ready to rise, and that combination usually means more pressure on ALTs even if Bitcoin only pulls back a few percent.
TOTALES has technically improved, but momentum is not broad enough for me to trust an alt season narrative yet. The setup I am watching most closely is whether BTC.D and stablecoin dominance can finally cool off. If they do not, then this market is still Bitcoin-led at best and vulnerable to a sharper reversal at worst, especially as we head toward May.
Macro is not helping me get complacent either. DXY is pushing higher, USDJPY is back near dangerous levels, the VIX has closed its long, and equities still look stretched after extended overbought runs. Oil jumped on geopolitical stress while gold and silver remain messy, so there is still plenty of cross-market risk sitting under the surface.
On the chart side, I covered strength and weakness across XMR, SKY, ATOM, POL, CHZ, PYTH, STRK, RENDER, PENGU, MON, TAO, NEAR, STABLE, and ZBCN, plus a few higher-risk names and non-crypto picks that are flashing interesting bottoming or exhaustion signals. There are still trades out there, but my bigger message today is simple: stay selective, respect resistance, and do not assume April strength means May will be easy.
This was maybe my favourite episode thus far, the prompt is so underrated. I’ve noticed especially with 4o and to some extent Sonnet 3.5 the models tend to be more literal in following instructions. While this tends to give more consistent and accurate results they definitely have lost some ability to infer what you mean if you’re a bit ambiguous with the prompt.
The LVM (large vision model) revolution is coming a little after the LLM (large language model) one, and will transform how we process images. But there’s an important difference between LVMs and LLMs:
- Internet text is similar enough to proprietary text documents that an LLM trained on internet text can understand your documents.
- But internet images – such as Instagram pictures – contain a lot of pictures of people, pets, landmarks, and everyday objects. Many practical vision applications (manufacturing, aerial imagery, life sciences, etc.) use images that look nothing like most internet images. So a generic LVM trained on internet images fares poorly at picking out the most salient features of images in many specialized domains.
That’s why domain specific LVMs – ones adapted to images of a particular domain (such as semiconductor manufacturing, or pathology) – do much better. At @LandingAI , by using ~100K unlabeled images to adapt an LVM to a specific domain, we see significantly improved results, for example where only 10-30% as much labeled data is now needed to achieve a certain level of performance.
For companies with large sets of images that look nothing like internet images, I think domain specific LVMs can be a way to unlock considerable value from their data. Dan Maloney and I share more details in the video.
Really nice touch on connecting and personalizing individual data stories! 👍
Spotify’s Head of Global Marketing Experience Explains Why This Year’s Wrapped Is the Realest Yet
https://t.co/FuC9RbHPKB
Podcasters - what if I told you could offer your pod to any listener around the world, in their own local language but still keep it in your own voice? That’s the pilot we’re launching @Spotify!
It’s called Voice Translation and using AI, translates podcasts episodes into alternate languages, all in the podcaster’s voice. It’s pretty insane. Take a look and let me know what you think!
https://t.co/wwUYsd3Mgo
Today, we’re announcing that @Amazon will invest up to $4 billion in Anthropic. The agreement is part of a broader collaboration to develop reliable and high-performing foundation models.
Introducing LoRA Roulette 🎲
Two custom models are loaded at random every refresh 🔄 - can you find a fun way to combine them? 🎨
▶️ https://t.co/bTxcgwCwG7
Unbearably sad to lose Luiz. He was behind so many of Google’s technical achievements, writing the book on our data centers, leading the design of our computing infrastructure, and so much more. He was at the top of his field, earning ACM’s highest honor in computer architecture.
Luiz saw beauty in everything, be it a warehouse architecture, a major 9 chord, or a wing of a hyacinth macaw. I’ll miss our chats about nature, music, and football, esp Brazil and Barcelona. RIP my friend.
https://t.co/cstVXVyj64
How economic trends affects tech products, from quickly rising to losing market presence.
Hopin, the struggling virtual conference unicorn, sells events and engagement units to RingCentral https://t.co/6335RvSqOE via @techcrunch
Today in The Download, our daily newsletter: The race to lead China’s autonomous driving market, and Kiva’s controversial changes. https://t.co/LNNmpKEfLD
Impressive. MetaGPT is about to reach 10,000 stars on Github.
It's a Multi-Agent Framework that can behave as an engineer, product manager, architect, project managers.
With a single line of text it can output the entire process of a software company along with carefully orchestrated SOPs:
▸ Data structures
▸ APIs
▸ Documents
▸ User stories
▸ Competitive analysis
▸ Requirements