A weird tax loophole with early stage startup investing that almost no one knows about:
If you invest in a small startup that has raised less than 1M in funding
And the business fails
You can write off up to 50K (single) or 100K (married) from your ORDINARY income aka salary
We're open-sourcing Hy-MT1.5-1.8B-1.25bit — a 440MB translation model that runs fully offline on your phone, supports 33 languages, and outperforms Google Translate.
At 1.8B parameters, it matches commercial translation APIs and 235B-scale models on standard benchmarks. By quantizing to 1.25-bit, memory drops from 3.3GB (FP16) to 440MB — 25% smaller and ~10% faster than prior 1.67-bit approaches, with no accuracy loss.
Covers 33 languages, 5 dialects, and 1,056 translation directions including minority languages like Tibetan and Mongolian.
Our translation model has won 30 first-place rankings in international MT competitions and is already deployed across multiple Tencent products.🏆
📲Demo APK (Android): https://t.co/DbcwLBe6vw
🤗Hugging Face:: https://t.co/NgXKeNZz4L
🔗GitHub: https://t.co/b6DqGKDelf
📄Paper: https://t.co/1a6repNrnt
Ernie-5.1 from @ErnieforDevs lands at #13 in Text Arena — now the #1 highest-ranked model from a Chinese lab.
Strongest categories:
- #9 Math
- #1 Legal & Government
- #4 Business, Management & Financial Ops
- #7 Software & IT Services
Congrats to the Baidu @ErnieforDevs team on this achievement!
Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data. Try it below!
with @AlecRad and @status_effects 🧵
China has decided to block Meta’s $2 billion acquisition of agentic AI startup Manus, making a surprise move to unwind a controversial deal that’s drawn fire for the leakage of technology to the US https://t.co/Ka9eQeAGXf
@nico_laqua I think it’s a mistake that people make early in their career to think that doing more stuff faster will get them farther.
There are much more powerful ways to build leverage.
SpaceXAI and @cursor_ai are now working closely together to create the world’s best coding and knowledge work AI.
The combination of Cursor’s leading product and distribution to expert software engineers with SpaceX’s million H100 equivalent Colossus training supercomputer will allow us to build the world’s most useful models.
Cursor has also given SpaceX the right to acquire Cursor later this year for $60 billion or pay $10 billion for our work together.
Tech used to celebrate being socially awkward if you were cracked at building stuff.
That’s dead. If you’re a pain to work with, you won’t have a spot at tech co’s of the future.