Dan Mac
@daniel_mac8
ai @ampcode | writing token stream | goodness, truth and ai | building at http://github.com/DannyMac180
ai agents and 'prompts' are history think not about prompts but about clearly defined 'cognitive destinations' openai and google's imo gold achievements just proved what i've been seeing with @AmpCode. most of us are thinking about using ai agents all wrong read in full 👇
occasionally I get the feeling that people on X don’t seem interested in ai or tech topics on a given day then I waltz on over to the ‘news’ tab and inevitably go: “Ooooooooohhhhhhh…..” 🤣
btw this method also works for your own mind it's becoming clear that all the wise ways to engage with ai are the same as the wise ways to engage with your own mind. we have a hard time seeing our own mind objectively because we are mostly completely identified with the mind…
ai agents and 'prompts' are history think not about prompts but about clearly defined 'cognitive destinations' openai and google's imo gold achievements just proved what i've been seeing with @AmpCode. most of us are thinking about using ai agents all wrong read in full 👇
THIS is the beauty of ai. today, anyone can harness the power of universal computation it's no longer the domain of wizards who spent years learning the secret incantations it's the domain of all humanity now.
Adrienne is the Executive Business Partner & Office Manager at @AmpCode She is self described as "non technical" Yet this weekend she build a chrome extension that saves her 1 day a month worth of work Here is her story: x.com/i/broadcasts/1…
White House set to announce AI action plan tomorrow here’s my question: has Donald Trump ever talked to an ai? which one did he use? what did they talk about???
We got an early account of what's in Trump's AI action plan, set to be announced tomorrow (link in next tweet) Subscribe to our @TIME newsletter In the Loop for reporting like this straight to your inbox 2x per week!
What did Sriram see?
Tomorrow should be a huge day for American AI. 🇺🇸
somehow reassuring to know that ai models can also get themselves all tied up in their cognitive underwear by overthinking
New Anthropic Research: “Inverse Scaling in Test-Time Compute” We found cases where longer reasoning leads to lower accuracy. Our findings suggest that naïve scaling of test-time compute may inadvertently reinforce problematic reasoning patterns. 🧵
💫 OpenAI Stargate when completed. (at least that's what my sources tell me)
we have signed a deal for an additional 4.5 gigawatts of capacity with oracle as part of stargate. easy to throw around numbers, but this is a _gigantic_ infrastructure project. some progress photos from abilene:
In this episode, Beyang and Thorsten discuss strategies for effective agentic coding, including the 101 of how it's different from coding with chat LLMs, the key constraint of the context window, how and where subagents can help, and the new oracle subagent which combines…
whoa 🤯 this thing is a BEAST.
>>> Qwen3-Coder is here! ✅ We’re releasing Qwen3-Coder-480B-A35B-Instruct, our most powerful open agentic code model to date. This 480B-parameter Mixture-of-Experts model (35B active) natively supports 256K context and scales to 1M context with extrapolation. It achieves…
the percentage of songs in which Robert Plant essentially has a vocal orgasm into the microphone is quite frankly unacceptable still some of my favorite music though
🎯 spot on. somewhat alarmingly so??? >> recommend trying this @diabrowser skill called /graham from Chris Paik
I’ve been evaluating people using my @diabrowser /skill. It’s called /graham and It’s scary good. Watch me test it on a few victims live (incl. @nikitabier )
X algo only permits me to go viral on the weekend @nikitabier pls fix. ty.
wish i could show you what people say in the DMs 😍
💯 We are trialing Amp with a few engineers. If their experience is the same as mine has been, we are going to roll it out more widely.
this morning my 2yo decided it was a good idea to run full speed into the back of the couch and smash his face against it. it didn't end well (he's fine it just obviously hurt) how many brain cells do we lose as young boys by being complete dumbasses for the fun of it? i don't…
if you are the type of person who’s highly competent but an absolute disaster to work with you are going to have a hard time.