one should seriously acknowledge how fast the world is changing and recalibrate their career plans especially cs/eng degree students - entry level tech jobs are becoming scarcer with layoffs + AI - many big tech companies not giving return offers to interns - swe’s who got laid

x.com/i/article/1810…

Watching the frenzy over Trump and Melania's crypto tokens is mind-blowing. Should I regret missing out or mourn the collective dive into memecoin madness? 💀

If Harry Potter was set in North Korea AI is wild

Since ghibli is trending and everyone is making ghibli version of real images. Here's a ghibli image made "real".

I think Elon Musk once said something like, 'The most ridiculous outcome is often the most likely.' I saw that in action today. My collaborator Vin Agarwal presented results on our QuantaQuill project at Biostate AI — a bioinformatics co-pilot that writes academic papers from

for the last 6 weeks we've built something wild. an AI that talks to your whole company. every team, every week. it sits in your meetings and reads your Jira boards. constantly identifying communication gaps, catching misalignment in real time before it turns into wasted

I have 9 VC meetings tomorrow. I can’t keep this up. If you’re a VC and interested in Sentra just call Sentra herself at (631) 562 6622 She will do the intro meeting for you!

You could have GPT-10 but unless it deeply understands who it's trying to serve, it will never be useful. This human <> AI context layer is a billion dollar problem. We're building the best way for an AI to really understand an organization. We find tribal data that isn't

"One time the [--] team deprecated an ad type without telling us. So we wasted a month building support for it even though it was deprecated." A month's salary for 8 devs is $100,000+. No OKR/retroactive performance monitoring software would've detected this in real time.

One of our biggest technical breakthroughs at MIT was when we solved distillation Incoming data is 99% junk. It's the reason why when you read a book, you forget most of the words Chunking + vector index dump is extremely crude, you need a way to filter the noise on ingestion