Last Week in Crypto - Oct 2
Join me in a fun little experiment
In the spirit of just doing things, I built a thing. I’m not sure if it is good or even useful but that’s ok. Putting it out there is the only way to find out. The thing in this case is an automated1 and algorithmic way to ingest crypto related news from across the entire internet2 and then assess what news was of significance3.
The problem I’m trying to solve4: I want to stay up to date with what is going on in crypto but I don’t want to spend hours sifting through the noise on Twitter,56 Linkedin or random news sites to find the signal.
A slightly more detailed version of how this works:
Every Thursday afternoon my little AI agent goes out and pulls in a wide swath of news stories it believes to be crypto related
It then reads all the stories and for each it writes a headline in plain english7 describing what the story is actually about
Next, it rates them on a 1 to 5 flame emoji scale on how significant they are
Finally, it reformats all of it in a friendly format I can copy and paste into substack
The plan is to run this for the next few weeks so I can show people and get feedback. If it’s helpful I’ll iterate and improve on it. If it’s bad I’ll just stop posting. I went with substack as the delivery mechanism because that was the fastest, lowest effort way to get it out there.
Some of the specific questions I want feedback on8:
Is this a real problem people face?
Does this effectively solve the problem?
What is the ideal delivery mechanism?
Are there significant stories it misses?
How accurate is the fire emoji ranking?
What else would you like to see happen?
I’ll also be able to measure open rates, subscribers9 and views so that will give a tiny bit of quantitative feedback. Hopefully this doesn’t turn out to be adding more AI slop to the universe.
Without further ado, I present Last Week in Crypto:
• Swift announces development of blockchain-powered ledger alongside 30+ banking institutions with initial prototype being constructed in collaboration with Consensys (via Bloomberg) 🔥🔥🔥🔥🔥
• Investment firm Vanguard, traditionally skeptical of digital currencies, readies to permit access to cryptocurrency ETFs through its brokerage service (via Bloomberg) 🔥🔥🔥🔥
• Tether announces plans to launch its fresh USAT token on Rumble platform, which will debut crypto wallet feature, targeting American market penetration; Tether maintains 48% ownership in Rumble (via Bloomberg) 🔥🔥🔥🔥
• Stripe introduces Open Issuance, a Bridge-supported platform enabling enterprises to mint proprietary stablecoins, utilized by digital wallet Phantom for launching its CASH token (via CoinDesk) 🔥🔥🔥🔥
• Flying Tulip, established by DeFi pioneer Andre Cronje with ambitions of constructing comprehensive onchain trading platform, secured $200M seed funding at $1B token assessment (via The Block) 🔥🔥🔥🔥
• White House pulls Brian Quintenz’s CFTC chair nomination following pressure from Winklevoss twins who urged President Trump in July to reconsider appointment (via Politico) 🔥🔥🔥
• Robinhood executive Vladimir Tenev reports Robinhood Prediction Markets surpassed 4B event contracts traded lifetime, with 2B+ occurring in Q3 exclusively; HOOD shares climbed 12.27% (via Bloomberg) 🔥🔥🔥
• European investigators examining whether Northern Data fraudulently claimed tax benefit on approximately €500M GPU inventory intended for AI but deployed for cryptocurrency mining (via Bloomberg) 🔥🔥🔥
• Financial advisory platform Wealthfront publicly files with US SEC for IPO, disclosing net earnings of $60.7M on $175.6M revenue during first half of 2025 (via Bloomberg) 🔥🔥🔥
• Chinese national admits guilt in London courtroom for laundering bitcoin connected to £5B Chinese investment fraud where UK authorities confiscated 61,000 bitcoin in 2018 (via Financial Times) 🔥🔥🔥
• Digital micro-lending company Tala reaches $340M annualized revenue while maintaining unprofitability 11 years after launch, planning to double lending operations by 2027 (via Forbes) 🔥🔥

Technically I type a command, hit enter and then wait about 45 minutes and then copy and paste the output here.
Ok obviously not the entire internet but a cross section which seems like it will reasonable have what I’m looking for. Scraping the web is a weird cat and mouse game right now. There is a high chance the entire thing just breaks at some point.
I’m realizing “significance” is in the eye of the beholder and AI is not particularly good at assessing it. Tweaking this part should be fun though.
Secondary problem: I was personally curious about this and if I could get it to work so I wanted to just make it.
Also I’m pretty sure prolonged exposure to Twitter will turn me into a Nazi which is something I’d like to avoid.
Also also I refuse to call it X. What do we call tweets now? Xeets? Yuck
I have a theory that so much of what is written on the internet can actually be described in one sentence but because of years of escalating SEO, everything is 100x more verbose than is necessary to convey the actually information.
Feel free to proactively answer these and let me know!
Hopefully this doesn’t cause an exodus of subscribers but sometimes that’s the cost of SCIENCE
