What’s happening with #Binance is crazy. Honestly this looks like a safer place than some hardware wallets.
Such transparency, preparation for a rainy day, and incident management is yet to be seen in this industry.
@cz_binance you and your team have earned a ton of respect.
12
Jonathan Barazany retweeted
Israelis are amazing. A bunch of them on a plane. They hear a ruckus in the cockpit, smash the door open, save the captain who has been stabbed by the co-pilot, subdue and neutralize him, then they commandeer the plane and land it. Everyone survives. Name any other nation in the world that produces the kind of people capable of pulling this off.
The end of ML?
Netflix replaced their 15 years old recommendation algorithm with an LLM.
It’s called "GenRec" and it completely changes how recommendation algorithms are built.
For over a decade, the Netflix recommendation engine was a masterclass in feature engineering. Data scientists built thousands of complex, handcrafted features to figure out what you wanted to watch next.
It required bespoke architectures. Massive infrastructure. Constant manual tuning.
Netflix threw all of it away.
They built GenRec, an LLM-backed ranker.
Instead of translating your behavior into complex math, they just turn your watch history and metadata into a natural language sentence.
They feed that raw text into a foundation LLM.
The AI simply reads your behavior like a story, understands your evolving tastes, and scores the entire catalog in a single forward pass.
No manual feature engineering. No complex bespoke architectures.
Here is the part that should terrify traditional data scientists.
This text-based LLM didn't just match the highly tuned production system Netflix spent years perfecting.
It beat it.
And it achieved those statistically significant gains using roughly 40x fewer labeled training examples.
We are watching a massive paradigm shift in real time.
The most complex predictive algorithms in the world are being replaced by models that just know how to read.
If Netflix can replace their core product engine with an LLM, what complex system in your business is about to become obsolete?
22
When Sonnet 5 was released, it came with a new tokenizer that generates roughly 30% more tokens for the same text. While Anthropic reduced its price from $3 to $2 to even that - they did not adjust the 200K context window that comes with, so effectively we have less memory now.
2
41
Jonathan Barazany retweeted
So jev is basically 2022 machine learning, without the need for a minimum training dataset.
Replying to @iam_zachi
@iam_zachi built a Postgres extension that lets you write this:
SELECT * FROM people
WHERE jev(people, 'the name sounds European')
No index. No embeddings. No vector column.
It works. Let me explain what's actually happening, because it's not what most people assume.
🧵
I think I just cooked something 🔥
jev(): a PostgreSQL extension that searches your whole database in natural language. No index, no embeddings, just one function.
WHERE jev(people, 'could work from home')
or
WHERE jev(people, 'name sounds european')
129 rows judged in ~1s for $0.0009. Second run: 6ms from cache.
2
1
102
Now the part the demo doesn't show.
A billion rows at ~160 tokens each = 160B tokens ≈ $6,700 per query.
25 million API requests. Weeks of wall clock.
The default prefetch cap is 5,000 rows. The public demo caps at 600. That tells you the intended scale.
1
8
Are we expected to get a newer Haiku version soon? @AnthropicAI @bcherny
19
Jonathan Barazany retweeted
הרבה ארגונים מדברים פתאום על ״אימון מודלים״ ומתכוונים לאימון מודלי שפה גדולים משל עצמם (ע״ג קוד פתוח סיני כמובן).
אם המשימה שלכם אינה ״יצירת טקסט״ (קוד, דוחות, שיחה וכו׳) - כנראה שהמודל שאתם רוצים לאמן הוא פשוט מודל קלסיפיקציה. כזה שמקבל טקסט ומחלץ ממנו ״ דברים״. מודלי שפה מצוינים בזה כשהבעיה לא מוגדרת, אבל ברגע שהיא כן, לאמן מודל עצום בשביל לבנות json של עשרה שדות מ-log נהיה די מיותר.
כן, זה פחות סקסי לאמן BERT מלעשות fine tune ל-Kimi אבל זה גם דורש פחות דאטא בשני סדרי גודל ויכול לעלות לפרודקשן מחר.






