Lance Hasson
@lancehasson
Followers
562
Following
18K
Media
123
Statuses
2K
Product @upwork. Prev founder @objective_inc (acq by Upwork), ML at the fruit company 🍎. I learned about computers @georgiatech. I like being outside ⛷🏄🚴♂️
Atlanta, GA
Joined April 2013
With agents becoming the default interaction model for apps, defining the right primitives for the product becomes incredibly important. LLMs can discover & dynamically compose any function you make available to them, thus the limit is whether the right atomic units exist
0
0
1
Yes. Also, keep the list short and stick to it. Only a few things really matter
The only to-dos method that has worked for me: Every week I make a weekly note, and write my to-dos for the week. I may add more items to it during the week. If any items didn't get done I roll them over to the next weekly note or drop them. That's it.
0
0
0
have to applaud whoever is running @eurofounder. top tier content 👏
0
0
0
anyone who hung around @CounterStrike in 2014 knows exactly how prediction markets on smaller sporting events will play out. a website called CSGO Lounge enabled unregulated gambling on matches. small teams made more money fixing matches until it blew up and was banned
0
0
2
This is the Will Smith eating spaghetti moment for world models
Introducing Marble by World Labs: a foundation for a spatially intelligent future. Create your world at https://t.co/V267VJu1H9
0
0
4
Oxford study finds LLMs are correct 94.9% of the time in tests, but humans in the same scenarios identified the correct conditions less than 34.5% of the time This is the age old problem of out of distribution data. The way users interact w/ chatbots is not "textbook"
1
0
2
Yupp launches to build a simple way to compare & evaluate LLMs. Evaluating LLMs in a real world setting is a huge problem, and is increasingly important with many static benchmarks saturating. Space is heating up with LMArena also raising $100m! https://t.co/rk4zM4uWLu
a16zcrypto.com
0
0
2
Very cool to see LMArena land significant investment and excited to follow their ambitions expanding their benchmarks. Evaluating LLMs (with agentic capabilities) by real users and in real world settings is extremely valuable work https://t.co/U5fC42fRvt
techcrunch.com
LM Arena, a crowdsourced benchmarking project major AI labs rely on to test and market their AI models, has raised $100 million in a seed funding round that values the organization at $600 million,...
0
0
0
Anyone with a B2B business with any type of success eventually learns this
0
0
0
Being grandfathered into FSD will be one of the biggest flexes
0
0
1
If you know model names you’re a 1% power user helping them test new capabilities ahead of taking them mass market
0
0
0
OpenAI model names don’t matter. 99% of users have no idea what model they’re using. They release new model features under new models with obscure names. When they get traction they upstream them into the main model used by ChatGPT by default
1
0
1
if you work with data check out Bauplan!
🚀 Introducing Bauplan A serverless, code-native platform for building data and AI pipelines — directly on your object store. No clusters. No notebooks. No GUI based workflows. Just Python + SQL + S3. 👉 https://t.co/aJGGCltlAg
0
0
1
What’s the positive angle on this?
Sources: the Trump administration plans to eliminate the IRS' free Direct File tax program, after IRS staff were told in mid-March to stop working on it (@fatimathefatima / Associated Press) https://t.co/GPIaco0SEJ
https://t.co/fBJa6KUJLl
https://t.co/ZOzeer2dpR
0
0
1
the most depressing tweet
I can't wait to see how much Nancy Pelosi made over the last 10 days
0
0
0
“Giving is one of those things.. it increases your quality of life, and it improves the world” 100% agree! Love that perspective represented
How much money do you need until it's enough? I ask this a lot on Moneywise and the answers are always interesting: - Bryan Johnson: $7M was his baseline number. Simple math: 7% interest vs inflation in rural Utah. Enough to "buy myself free time" with basics covered. - Vinay
0
0
1
Every time I see a chart like this I’m reminded how nearly everything follows a normal distribution
How bad was today's stock market decline? Of the 11,406 trading days since January 1980, only 29 had larger declines.
0
0
9