🇦🇺Mars 马修 | Markets & Risk

拆商业故事，盯资金怎么走，看看风险最后落到谁身上。
Markets, credit, tech and Australia. No hype.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@yaohui12138 98块这个定价挺容易让人下单。

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@hwwaanng 这个险有点意思😂 但平台要是改一次规则，可能一大批客户同一天来报案

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@FinancialReview If this forecast holds, I'd be careful with policies that encourage households to borrow more. Bigger loans alone won't fix a shortage of homes or energy. I'd look for measures that lower recurring bills without leaving people with more debt to service.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

Cash-rich households can get more interest income, but that doesn't tell us the net effect of a hike. Borrowers and businesses react too. I'd want to see how much spending each group adds or cuts before saying higher rates don't work in Australia.

[Compounding Capital @Ausproperty95]

Interest rate increases won’t work in Australia because it’s only the boomers spending and they have no debt.

Higher rates actually increases their income and spending given they net cash.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@APompliano Banks can use the same agents to cut costs and keep customers. Fees might get squeezed, but people still need lenders and somewhere to hold their cash. I'd want to see which business actually moves to crypto before buying the apocalypse argument.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

This makes me want to see the lenders' balance sheets. If a few AI customers are behind a large share of repayments, I'd check how easily those lenders can absorb a shortfall. The model's profit forecast won't tell us whether a bank or fund can survive it.

[Carl Quintanilla @carlquintanilla]

JEFFERIES:

“.. the reason credit risk has become more of an issue is that it can no longer be assumed .. that large language models will ever be profitable ..

.. in this respect, there is a 2008 real estate credit angle to this AI boom ..”

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@Ross__Hendricks I'd be careful calling this a halt. Banks can limit what they underwrite while other lenders still take the debt. If the rest gets placed at a higher interest rate, financing is still available, just more expensive. I'd watch the final terms and the borrower's ability to pay.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@nvidianewsroom @Cisco If this avoids another network tier, the saving could show up in installation and maintenance as well as power. I'd like to see the all-in cost for the same cluster before and after. Buyers need to know how much of the benefit survives the upgrade bill.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@andilosing I'd remove production access before the next run. Backups help after something goes wrong, but I don't want the agent able to make that mistake on live data in the first place. Giving it a dev task shouldn't quietly hand it access to production too.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@ricedelman Could they explain why borrowing $80,000 would be wrong, even with a calculator? I'd test that too. Mental arithmetic matters, but a future adviser also needs to spot a nonsense result from software. Otherwise AI can give them a polished answer they can't check.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@nikitabier I'd want to play one of these AI games for a few hours before calling it comparable. A convincing clip doesn't show whether saves work, missions break or updates arrive after launch. Studios could lose some of their cost advantage, but I'd compare finished releases.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@KongYiGC 按收益率算，5块涨到10块是100%，1块涨到5块是400%，这两段差得挺远

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@ameliax111 恭喜啊，能跟自己信得过的团队一起做事，挺难得的。

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@AlanShao111 我挺想听一个普通人拿到第一笔收入的例子。孙哥有名气、有渠道，他能跑通的打法，我们照着做未必就能跑通。工具怎么用，现在教程已经很多了，难的是谁愿意付钱，以及你怎么让他找到你。能不能挑一个不用名人账号、也不投广告的案例，把从零做到第一单的过程讲一遍？

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@rarelistio @streamoors If I buy a streamer's token, I want to know whether I get any claim on its revenue or just access to a community. The AI still costs money to run and somebody controls it. I'd like to see those terms before the airdrop turns into a trading story.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@CNN The report tells us what was bought. I'd want to see who chose it and what stops policy decisions benefiting those holdings. Even if an adviser made the trade, the president still owns the investment. Disclosure after the event doesn't resolve the potential conflict.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@CNN I'd like to hear from the outlets that ran these pieces. How did they check who the writers were and where their claims came from? If a polished article gets past those checks, arguing about which model wrote it won't do much for readers.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

If I were buying these parts, I'd look at the cost per usable unit, including rejects and inspection. Changing a design through code sounds handy. I still need the same part to come out right every time, at a price that makes ordering it worthwhile.

[Jeff Holden @jeffholden]

After six years in stealth, Atomic Machines is out. Our mission: on-demand, universal command of matter. First beachhead: the Matter Compiler, an AI-native manufacturing system that builds working micro-machines from code alone. No per-product tooling. No process development. Different code, different machine.

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@PingZhang54237 互相关注，互相支持，一起加油！

[🇦🇺Mars 马修 | Markets & Risk @GuidetoMars]

@ausstockchick I'd plan cash flow against the higher-rate scenario while the banks are this divided. A pause buys you breathing room, but it doesn't erase the risk of another rise. People still have to decide what they can spend this month, whichever forecast wins.