Mintttch🐣

Now starting DCA with stocks, crypto, and finding other things to do along the way Feeling like I'm losing passion for crypto after being in it for 4-5 years . Starting to get tired, and at the end of this year, I'm moving to the countryside after living in Bangkok for almost 10 years. Want to live a slow life 😅

Anthropic agrees to pay SpaceX about $1.25 billion per month for compute usage from the Colossus data center . This deal will run until May 2029 with Anthropic using compute from both the Colossus data center in Tennessee and SpaceX's new Colossus 2 data center . SpaceX's S-1 filing states Anthropic will receive special rates in May and June and either party can cancel the contract with 90 days' notice . If this deal runs through the full term SpaceX will earn more than $40 billion from Anthropic . An Anthropic spokesperson confirmed the $1.25 billion per month figure to Business Insider with this compute set used for inference or running AI models to serve customers . The S-1 document also states SpaceX expects to make similar deals in the future viewing this as a way to monetize unused compute capacity . However, SpaceX's AI infrastructure costs are very high The document notes losses from AI operations last year rose to over $6 billion and nearly $2.5 billion loss in the first quarter this year . The main causes are cloud costs, GPU expenses, and chip depreciation . Interestingly, SpaceX also states in the document one of their major investment plans is to produce their own GPUs . If realized, SpaceX could move to compete with NVIDIA which is currently a key player in the advanced GPU market . In the S-1, SpaceX also categorizes AI Infrastructure as a key market and estimates this market could be worth around $2.4 trillion

KitKat is so expensive nowadays, haven't bought it for a while Just checked again, it's 35-45 baht now? . A few years ago, it was still 20 baht, wasn't it?

Verus-Ethereum Bridge was attacked It was reported that assets worth about $11.58 million or approximately 378 million baht . The main assets drained were about 1,625 ETH 103.57 tBTC and 147,000 USDC . An interesting point is that Verus had communicated that their bridge has no smart contract risk because it uses protocol-level validation, cryptographic proofs, and notary witnesses . But from the preliminary analysis The problem was not due to signature hacking Not due to stolen notary keys And not a parser bug . But a gap in verifying whether the value to be paid out on Ethereum is actually backed by assets on the source side . Simply put, the bridge correctly checks many things but missed the most critical point that "Is the money being paid out truly backed?" . The attacker’s cost was identified as only a few dollars in fees but they were able to extract assets worth over $11 million . And this is not a small case in the overall DeFi landscape . According to PeckShield data In 2026, crypto bridge exploits caused total damages of about $328.6 million from 8 major incidents . This means bridges remain one of the most vulnerable points in crypto because they have to connect chains, liquidity, proofs, validators, contracts, and multiple verification systems together . This case reiterates once again that cross-chain bridges are still a major risk point in DeFi

Yesterday @Figure_robot held a live event between a “human” and a “robot”. . The challenge was to sort parcels on a continuous conveyor belt for 10 hours. . - The human side was Aime, an intern kid - The robot side was Figure 03 or F.03 named Bob working fully automatically with AI Helix-02. . The task seemed very simple which was to look at the barcode → pick up the parcel → place it on the conveyor belt with the barcode facing down. . But it had to be repeated thousands of times, so it wasn’t easy at all. . In the end, the human still won - Aime managed 12,924 pieces (averaging 2.79 seconds per piece) - Bob managed 12,732 pieces (averaging 2.83 seconds per piece) . Only a difference of 192 pieces after competing for 10 hours. . And although Aime won, after the competition she said her left arm was almost broken, her fingers blistered, and she was very tired. . Meanwhile, the robot could keep working continuously - No need to take lunch breaks - No need to use the bathroom - No fatigue - And no injuries (Actually, it’s not that great; it still makes mistakes often or throws things off the side sometimes.) . Brett Adcock, CEO of Figure AI, even posted somewhat boastfully, “This is the last time a human will ever win.” . Meaning, this might be the last time a human wins because the next generation of robots will be faster, tougher, and able to work continuously better. . In summary, today humans still win but it’s a very close victory (not sure if they tried to make it close because humans have to take lunch and bathroom breaks, etc.)

Jamie Dimon, CEO of JPMorgan, believes that AI will greatly improve people's lives. . He didn't just say that AI will take people's jobs, but said that AI might improve human life on a whole different level. . From his perspective, in the next 30 years, our descendants might only work 3 and a half days a week, live up to 100 years old, and many diseases that are hard to treat today might no longer be a major problem. . He gave examples that AI might help cure cancer, discover new materials, reduce road accidents, and make many industries much more efficient. . The phrase he said was “Life will be better.” . But what's interesting is that Dimon doesn't see AI in an overly optimistic way only. . He said the real risk is not just AI getting better, but AI changing the world too fast. . If productivity surges too quickly, some types of jobs might disappear before people can adapt. . And this is the point where he thinks the public and private sectors must seriously cooperate, not just let people lose jobs and then tell them to “go learn on their own.” . What should happen is those affected must be guided to new jobs, there must be reskilling, there must be labor shifts to industries still lacking workers, such as advanced manufacturing or other key industries. . This perspective is very interesting because it shows that AI is not just a technology issue. . But it is about managing the social transition. . In the long term, AI might really improve life, people might work less, be healthier, and the economy be more efficient. . But in the short term, many people might have to face the pain of changing jobs, learning new skills, or accepting that the jobs that were once stable might no longer be stable. . In simple summary, AI might make the world better. . But whether the world gets better for everyone depends on how well we help people adapt in time.

AI Future Ready thanks all Media & Community Partners ✨ Thank you to every community and partner who supports making AI learning more accessible for everyone Let's Learn • Do • Apply together 🙌🏻 📍 SCBX NEXT TECH 🗓 21 May | 13.00–18.00 🎟️ #AiLearning