Fascinated by Machine Learning and Computer Vision. I also run the SF-Machine Learning meetup: https://nitter.cf/t.co/HmEJuvzHYN
San Francisco, Ca
Joined May 2009
- Tweets1.5K
- Following2K
- Followers638
- Likes16.8K
Tony T retweeted
The American Dream starts with opportunity. 🇺🇸
Susan and I want to help jumpstart that future for the next generation. We are contributing $250 to the investment accounts of 25 million children ages 2–10 living in zip codes with a median income of $150K or less.
It takes just minutes to sign up and secure your child’s financial future.📈
You must sign up your kids, and now it could not be easier: trumpaccounts.gov/form
Tony T retweeted
"Prognosticative pastry." "A hound circling a tree, nose to bark."
These aren’t parodies - they’re actual quotes from SOTA models in response to creative writing prompts, and they’re winning leaderboards that are rewarding slop.
We’re introducing *Hemingway-bench*, a new AI writing leaderboard, to fix this:
surgehq.ai/leaderboard
surgehq.ai/blog/hemingway-be…
We designed Hemingway-bench to push frontier model writing toward genuine nuance and impact.
Instead of autograders and two-second vibe checks - both of which love fancy literary devices and dense formatting, over actual quality - we used expert human writers across a variety of fields to judge real-world writing tasks.
Why? I love writing. I love reading. Great science fiction is one of the things that's always inspired me. Even in terms of "enterprise value", so much of what we do in our day-to-day involves writing - we want crisp emails and insightful reports, not dry, verbose summaries.
Yeah, coding is important - but there's a reason I use CC-assisted apps, but still haven't read a full-fledged AI novel.
What did we find? Current leaderboards are easily hacked, and often negatively correlated with actual quality. If a model (over)uses all the stuff you learn about in school (metaphors in every sentence! transition words! complex, flowery phrases!), it ranks high on EQ-bench and LMArena.
But that’s not good writing that people actually want.
The winners of Hemingway-bench didn't sound like they were trying to win a poetry slam. Gemini 3 Flash, Pro, and Opus 4.5 took the top 3 spots because they had natural voices that didn't sound pretentious.
They were poetic and immersive, but in the right ways.
When they used wit, they didn't sound cringey and try-hard - they sounded like your naturally funny friend.
I'm waiting for the day AI wins a Pulitzer, and hopefully Hemingway-bench helps guide it on its way.
Check out the leaderboard and examples here: surgehq.ai/leaderboard
And our blog post describing it: surgehq.ai/blog/hemingway-be…
Tony T retweeted
I can’t believe this is real.
Gauntlet AI is taking over former adtech giant Marin Software out of bankruptcy.
We’re going to try to turn the company around in SIX WEEKS using AI.
With a reality TV crew filming the whole thing.
To follow along: Marin.GauntletAI.com
Tony T retweeted
The 3rd edition of my book Deep Learning with Python is being printed right now, and will be in bookstores within 2 weeks. You can order it now from Amazon or from Manning.
This time, we're also releasing the whole thing as a 100% free website.
I don't care if it reduces book sales, I think it's the best deep learning intro around, and more people should be able to read it.
"The proper function of man is to live, not to exist. I shall not waste my days in trying to prolong them. I shall use my time." - Jack London
Tony T retweeted
Shower of thoughts: Instead of keeping your Twitter/𝕏 payout, direct it towards a "PayoutChallenge" of your choosing - anything you want more of in the world!
Here is mine for this round, combining my last 3 payouts of $5478.51:
It is imperative that humanity not fall while AI ascends. Humanity has to continue to rise, become better alongside. Create something that is specifically designed to uplift team human. Definition intentionally left a bit vague to keep some entropy around people's interpretation, but imo examples include:
- Any piece of software that aids explanation, visualization, memorization, inspiration, understanding, coordination, etc...
- It doesn't have to be too lofty, e.g. it can be a specific educational article/video explaining something some other people could benefit from or that you have unique knowledge of.
- Prompts/agents for explanation, e.g. along the lines of recently released ChatGPT study mode.
- Related works of art
This challenge will run for 2 weeks until Aug 17th EOD PST. Submit your contribution as a reply. It has to be something that was uniquely created for this challenge and would not exist otherwise. Criteria includes execution, leverage, novelty, inspiration, aesthetics, amusement. People can upvote submissions by liking, this "people's choice" will also be a factor. I will decide the winner on Aug 17th and send $5478.51 :)
A great reminder from @patrickc that we can, and should, operate with urgency in life patrickcollison.com/fast
DPI seems to be pushing VCs to offload top startups in secondary deals. Smart move? #VentureCapital #incentives
DPI seems to be an important metric these days. DPI = Distributions/Paid-in Capital. In other words, the ratio of money returned to you divided by how much money you put in. If you invested in an early stage company, your DPI may be 0 for a while (or forever).
If you’re a startup employee that wants to learn more about how “the VC side of things” works check out @HustleFundVC . They put out great content and have interesting deals as well. It will help you better understand why some decisions are made at a company level as well.
One of the most useful tactics I’ve found when managing people is to establish a “How to best work with me” document inspired by @chughesjohnson (growth.eladgil.com/book/the-…). It helps set clear expectations upfront.
Tony T retweeted
So let me get this straight: investors in a 3rd-layer SPV will likely have to overcome three 20% carry hurdles (and not to mention other possible fees for SPV setup and management fees for EACH SPV) before they get paid?
That is, if they get paid if/when Anthropic exceeds $130B in valuation upon liquidity event?
This is straight up horrific. Not investment advice, but don’t do this!
Working with developers in Argentina you come to realize that there's a whole different form of Spanish conjugation that's not taught in American Schools (the "vos" form)🇦🇷
nitter.cf/i/grok/share/ntStGYl1Y…
Tony T retweeted
a general reminder to not trust any software that you didn't read or audit yourself
This quoted post is unavailable.
Tony T retweeted
C/C++ is still bread and butter of system engineering.
Even most models can be trained and exported via Python, most kernel drivers for the robotic parts are still written in C/C++ down to something as basic as i2c busses and DMA controllers.
You still need to be able to express your thought process directly to the machine how you want to actuate, with all the fault handling and hardening.
Onboarding a new Engineer to your startup? Quickest way to ramp them up is to have them act as a QA Engineer for the first 2-3 weeks.