- Enzyme
- Posts
- what a time to be AlIve
what a time to be AlIve
open source FTW, ETH ETFs approved, OpenAI attack on search and more
was this email forwarded to you? sign up here.
gm gm,
welcome back to another week of heated wars in AI over open-source models, and a major win for institutional crypto fans.
In today’s edition read more on:
Spot ETH ETFs get the green light
Meta goes all-in on open-source AI models
Latest news across web3 and AI featuring OpenAI and Mistral new models, MoonPay partners with Polymarket, the first app on the App Store for transferring memes and more.
As always, we’re using this space to provide you with the latest product updates and relevant industry news in the shortest form possible.
Let’s go!
IN THE SPOTLIGHT (1/2)
Spot ETH ETFs get the green light!
Just six months after launching the first spot BTC ETFs, the digital asset world celebrated another major win as Ether began trading on Tuesday.
It’s official: Spot Eth ETFs have been made effective by the SEC. The 424(b) forms are rolling in now, the last step = all systems go for tomorrow’s 930am launch. Game on.
— Eric Balchunas (@EricBalchunas)
8:48 PM • Jul 22, 2024
Eight issuers announced their fees, ranging between 0.19% and 0.25%. They are:
Grayscale Ethereum Mini Trust and Grayscale Ethereum Trust
Franklin Ethereum Trust
VanEck Ethereum Trust
Bitwise Ethereum ETF
21Share Core Ethereum ETF
Invesco Galaxy Ethereum ETF
iShare Ethereum Trust ETF
Fidelity Ethereum Fund
Ethereum welcomed its first day of trading with a relatively stable price, showing no major dips or surges, contrary to some market predictions of reaching around $5K.
On its first day of trading, Ether saw total net flows of +$106.6 million and $1.1 billion in total volume, which is 23% of Bitcoin’s first-day trading volume.
Yet, it was not an easy road.
After the initial approval in May, the issuers went back and forth with their forms until they were finally approved.
Solana next? Issuers are already submitting filings.
Exciting times are ahead for crypto vs. institutional wins!
IN THE SPOTLIGHT (2/2)
AI open source FTW!
Even big names can call it a week when it comes to AI.
Both Meta and Apple have released their open-source AI models. Let’s check out what each has to offer now.
First off, Meta has released Llama 3.1, now the largest open-source model with 405 billion parameters. It’s trained on 15 trillion tokens of data up to 2024 and features a context window of 128,000 tokens. The model comes in three versions based on the number of parameters: 405B, 70B, and 8B.
Starting today, open source is leading the way. Introducing Llama 3.1: Our most capable models yet.
Today we’re releasing a collection of new Llama 3.1 models including our long awaited 405B. These models deliver improved reasoning capabilities, a larger 128K token context… x.com/i/web/status/1…
— AI at Meta (@AIatMeta)
3:10 PM • Jul 23, 2024
The model handles tasks related to coding, math, complex reasoning, and multilingual summarization. According to the company, it outperforms GPT-4 and Anthropic’s Claude 3.5 Sonnet on several benchmarks.
Zuck is an advocate for open-source AI, believing it will follow a pattern similar to Linux versus other operating systems in the past. He adds that companies like Amazon, NVIDIA, AWS, Google, Dell, Deloitte, and others are joining Meta to expand the broader ecosystem.
According to him, this open-source ecosystem will offer greater benefits to humanity than keeping AI closed, highlighting advantages for the U.S., the economy, developers, startups, and others, with safety as a key priority. He asserts that open-source AI is the way forward.
In the meantime, Meta continues to expand its AI assistant to more languages and countries, now adding a feature for image generation based on a person’s likeness.
Apple also announced new developments, with its research team releasing a family of open DCLM models on Hugging Face.
We have released our DCLM models on huggingface! To our knowledge these are by far the best performing truly open-source models (open data, open weight models, open training code) 1/5
— Vaishaal Shankar (@Vaishaal)
3:18 PM • Jul 18, 2024
These models feature 7 billion and 1.4 billion parameters, outperforming Mistral and approaching Llama 3 in performance.
As part of broader research, these models likely won’t be integrated into Apple devices in their current form, but they are available for anyone to use or adapt.
What a time to be AlIve!
SUBSTRATES
What else is going on?
WEB3
Proton debuts Proton Wallet, an open-source, self-custodial wallet that allows users to transfer BTC via email addresses.
Moonshot is the first app on the App Store that allows trading memes with Apple Pay.
Ryan Selkis steps aside as Messari’s CEO to serve as a Senior Advisor due to personal politics and views expressed on social media.
The Solana Name Service (SNS), formerly Bonfida, has undergone a major rebrand to start a new era for .sol domains.
MoonPay partners with Polymarket, enabling users to buy crypto using debit/credit cards, Apple Pay, Google Pay, and PayPal.
Ferrari extends its acceptance of cryptocurrency payments to Europe, following the launch in the US.
Crypto.com celebrates its 8th anniversary with the BTC logo displayed on the Sphere in Vegas.
AI
Elon Musk announced the start of the ‘world’s most powerful AI’ training cluster.
OpenAI is testing SearchGPT, an AI search engine prototype offering fast and timely answers with clear sources.
Adobe has launched a new set of AI features for Photoshop and Illustrator.
Mistral announces Large 2, a 123B-parameter model, under a permissive research license.
SingularityNET invests $53 million in an AGI supercomputer to accelerate progress in AI research and provide resources for developers.
Lenovo plans to integrate AI into all its computers by 2027.
FROM THE REPOSITORY
Check this out!
Use now or save for later:
Archie will help you plan and design your next software app with the power of AI.
GombleGames protocol transforms fun and engaging web3 games into gamified web3 experiences with all aspects covered.
Build your online presence with StoriAI.
*Shipping amazing products or services in AI or web3? Reach out and get featured for free to your community of over 1k readers.
Visual of the week: Mainstream adoption TBD
SCOOPS OF THE WEEK
I made the closed-source vs. open-weight models figure for this moment.
— Maxime Labonne @ ICML (@maximelabonne)
7:12 AM • Jul 24, 2024
Do you know what form of currency hasn’t been affected by widespread cyber outages? Bitcoin.
Vires in Numeris
— Senator Cynthia Lummis (@SenLummis)
3:30 PM • Jul 19, 2024
Polymarket is the news.
It priced in a 10%+ chance of Biden dropping out, even before the debate. Meanwhile, legacy media claimed Biden's senility was a conspiracy theory.
No journos, only crypto.
— Balaji (@balajis)
9:19 PM • Jul 21, 2024
POV: your friend just asked you to explain crypto
— gaut (@0xgaut)
1:46 PM • Jul 21, 2024
And that’s it for the week - not so short but sweet.
Keep on building!
Got any questions?
Drop us a reply to this email.
Ready to build in web3?
Deploy in seconds - we’ve got smart contracts for pretty much anything!
Stuck or need a second web3 brain?
Get in on a free call with us.
New to web3?
Grab the web3 starter pack🎁 an essential guide for everything web3.
Reply