please wait while the wizard installs the software
b'65537
Joined August 2023
- Tweets558
- Following1.9K
- Followers308
- Likes74.5K
s retweeted
He replied, “I will take refuge on a mountain, which will protect me from the water.” Noah cried, “Today no one is protected from Allah’s decree except those to whom He shows mercy!” And the waves came between them, and his son was among the drowned.
11:43
The benefits of abundant intelligence will accrue far outside of San Fransisco.
Spent the last year writing about capital, sovereignty, and the coming reordering of economic power.
THE COMMANDING HEIGHTS
Commanding-Heights.com
s retweeted
ALT int getRandomNumber() { return 4; // chosen by fair dice roll. // guaranteed to be random. }
s retweeted
The cheapest model with data access beats the best model without it
We ran 9 models from 5 labs on 82 work tasks where the answer has to come from real data. Each model ran the same tasks with and without access to paid data.
Claude Haiku 4.5 is one of the cheapest closed-source models you can run an agent on. Claude Opus 5.5 is one of the most capable (and expensive). Haiku with data finished 60% of the tasks. Opus without it finished 50%. Per finished task, the cost came out about the same.
It wasn't just Haiku. Every model with data beat the best model without it, open source included. The weakest, Grok 4.7, still finished 54%. That edge over Opus is small enough to be noise, but the direction never changed across models or task categories.
Most of the agent conversation is about which model to use. On this kind of work, what the agent could access mattered more.
Method, tasks and traces in the replies.
s retweeted
Agents can now set up your website’s security with Turnstile Spin
It's surreal seeing a string of ideas and iteration turn into a blog like this! I'm so grateful for my time at Cloudflare :)
blog.cloudflare.com/turnstil…
s retweeted
There are 5-6 secret additional hours between 1 and 3 am if you know how to do it correctly. This will take years off your life and make you look and feel like shit but also it is peak Going On The Computer Time
s retweeted
Once you completely unmoor your sleep schedule from the rising and setting of the sun, you can access new, formerly grayed out sections of space-time, sort of the dust in between the stars, and most crucially, use it to cram 48 to 96 hours worth of screen time into one single day
s retweeted
Finding an API that actually fits your needs is hard.
There are so many similar ones out there that teams burn hours just researching which one to use.
We realized that was a problem worth solving.
Locus Pro is now substantially better at exactly that. You and your agents find the best tool for that specific job using plain natural language. Locus reads what each endpoint actually returns to make sure it fits.
s retweeted
the general consensus seems to be that coding is basically solved.
yet remarkably it doesn’t seem to have produced more than incremental improvements to most of the software we use.
coding is solved, they say, yet we are still using google drive and notion. coding is solved, they say, yet we can think of no better way to coordinate tasks than linear / jira. coding is solved, they say, yet we are still using gmail and outlook. and we still just message each other about everything in slack.
where is the new generation of tools? why is nothing changing?
why are these companies, supposedly in possession of this technology which has ‘solved coding’, hardly producing anything better than what they were creating before?
the great paradox of the moment.
s retweeted
We’re giving agents more tools! Now, any business with a ZeroClick-powered Agent Storefront is available as a verified service on Locus Pro.
If you have a product that you want to sell to agents, set up your own Agent Storefront with @zeroclick
Try it today!
s retweeted
Beware people obsessed with outcomes instead of building outcome machines. Its worse than ever with AI, but these people existed before. Short term results above all else, etc. Don't fall into the trap. Invest in building strong fundamentals, invisible supports, and outcomes flow like water. An outcome machine.
s retweeted
Apartment hunting in SF blows.
We needed to find one for 3 members of our team and everything is either off-market, 5x more expensive then we want to pay, or in a bad area.
But thanks to Locus Pro it took under 2min.
We pulled live data from RentCast, Zillow, Craigslist, and AirBnB through our agent; filtered by the neighborhoods we wanted at our price range; and called the landlords.
s retweeted
We just killed every single web search API you're using.
We are excited to announce our partnership with parse[dot]bot to let agents use any site on the internet via API.
Parse turns any website into a typed API and Locus puts all of the endpoints on one prepaid balance with our other providers.
Travel, e-commerce, finance and more.
Turn on the ones you want, and let your agents or platform call them by use.
Live now!
s retweeted
We just shipped self-serve for Locus Pro, the first OpenRouter for everything.
You shouldn’t need a dozen subscriptions just to make your agent useful.
Get access to 3,300+ providers and 16,000+ agent tools all through one prepaid balance.
Sign up, paste in the MCP URL, turn on the tools you want. Your agent can use OpenAI, Anthropic, Firecrawl, ElevenLabs, Exa and thousands more without you opening a single new account.
No subscriptions or monthly minimum, top up with as little as $5.
s retweeted
I was once part of an effort to predict defect counts from metrics like function length, cyclomatic complexity, object hierarchy depth, and so on.
The key finding: every one of them was just a proxy for lines of code. More code, more bugs. No need to overthink it.
Counterintuitively, comments counted too. A heavily commented file had more defects, not fewer. Which really makes me hate LLM love of comments even more.
Corollary: verbose languages mean more defects per same functionality.
