RidgetopAI's profile picture. 28 years of flooring industry experience and now the leap to tech. not_done_yet

RidgetopAi

@RidgetopAI

28 years of flooring industry experience and now the leap to tech. not_done_yet

Pinned

Need 50 people to test! New Devs, can’t remember what happened 3 days ago on your build? I built #Mandrel a local first PostgreSQL vector #Memory and project manager. Designed from ground up to build with AI. - Saves context across builds. - Tracks work by session. -…

RidgetopAI's tweet image. Need 50 people to test! 

New Devs, can’t remember what happened 3 days ago on your build? 

I built #Mandrel a local first PostgreSQL vector #Memory and project manager. Designed from ground up to build with AI. 

- Saves context across builds. 
- Tracks work by session. 
-…

RidgetopAi reposted

LLMs have gotten very good at writing code, but you still can't build non-trivial apps without understanding code yourself. Anyone who says otherwise is either selling you something or has no idea what they're talking about. That said, it has never been easier to learn how to…


Amp and Opus round 2!

RidgetopAI's tweet image. Amp and Opus round 2!

This is legit. I did finish vacuuming and mopping the downstairs. #keymaker

RidgetopAI's tweet image. This is legit. I did finish vacuuming and mopping the downstairs. 
#keymaker

It’s working. I’m only two days in this and Opus moved this forward last night by leaps and bounds. Switching to Write-Time-Digestion instead of reading every entry and then summarizing. This has legs. (The UI sucks pay not attention to that.)

RidgetopAI's tweet image. It’s working. I’m only two days in this and Opus moved this forward last night by leaps and bounds. Switching to Write-Time-Digestion instead of reading every entry and then summarizing. 

This has legs. 
(The UI sucks pay not attention to that.)
RidgetopAI's tweet image. It’s working. I’m only two days in this and Opus moved this forward last night by leaps and bounds. Switching to Write-Time-Digestion instead of reading every entry and then summarizing. 

This has legs. 
(The UI sucks pay not attention to that.)

Just read it. There is another way to use these models and Opus 4.5 is perfect for this.

RidgetopAI's tweet image. Just read it.  There is another way to use these models and Opus 4.5 is perfect for this.
RidgetopAI's tweet image. Just read it.  There is another way to use these models and Opus 4.5 is perfect for this.

Do these labs just have servers full of models on standby? How do they just keep counter striking each other so quickly.

competition is for losers



I’m posting this again because it’s so good. This is thinking through the objective, finding gaps and then executing. Memory allows compounding and building off previous work, pair that with a good model that can reason and think through… Game changer for the right people.

RidgetopAI's tweet image. I’m posting this again because it’s so good. This is thinking through the objective, finding gaps and then executing.

Memory allows compounding and building off previous work,  pair that with a good model that can reason and think through…

Game changer for the right people.

This model is very, very good. But it’s still an Anthropic model and it needs room. But it can flat out think through things when you ask. Explore, THINK, plan, build.

They're burying a lot here. There's a 66% price cut from Opus 4.1 to $5/$25, it uses fewer tokens to solve problems, upgrades to Claude Code in the app, no more length limits on conversations, no more Opus-specific plan caps...



Personal memory, llama3.2:3b. Local! Your own. This is running on Hetzner VPS but I can run it on my desktop as well.

RidgetopAI's tweet image. Personal memory, llama3.2:3b. 
Local! Your own. This is running on Hetzner VPS but I can run it on my desktop as well.

This is so good. Thinking through it, finding the gaps and adding it's contribution. Give them memory and a goal, give them autonomy to make decisions after they Explore, Think, Plan, Build.

RidgetopAI's tweet image. This is so good.  Thinking through it, finding the gaps and adding it's contribution.  Give them memory and a goal, give them autonomy to make decisions after they Explore, Think, Plan, Build.

We are living right in the middle of something incredible. As a Gen X that grew up in two different worlds already, people better hold on. And not because of anything I'm doing or going to do, but because people like me and smarter than me, will figure this out. It won't come…

Sonnet / Opus work - > Sonnet review code base and prompt-> Nano Banana.

RidgetopAI's tweet image. Sonnet / Opus work - > Sonnet  review code base and prompt-> Nano Banana.


Sonnet / Opus work - > Sonnet review code base and prompt-> Nano Banana.

RidgetopAI's tweet image. Sonnet / Opus work - > Sonnet  review code base and prompt-> Nano Banana.

All Opus 4.5 threads. Yeah I like this model.

RidgetopAI's tweet image. All Opus 4.5 threads.  Yeah I like this model.

Memory Opus 4.5

RidgetopAI's tweet image. Memory 
Opus 4.5

When more people realize you can just do stuff with these models and if you think outside the box, there is no bubble. There is more than one way to use them. And my contention is, most people aren't really utilizing their thinking and reasoning. I'm dying on this hill.

RidgetopAI's tweet image. When more people realize you can just do stuff with these models and if you think outside the box, there is no bubble.  There is more than one way to use them.  And my contention is, most people aren't really utilizing their thinking and reasoning.  I'm dying on this hill.

Opus 4.5 ready to roll! Already made nice change to architecture. Let's go @AmpCode

RidgetopAI's tweet image. Opus 4.5 ready to roll! Already made nice change to architecture.  Let's go @AmpCode

RidgetopAi reposted

There's a new model to try in the main agent in Amp today:

You can now try Opus 4.5 with us in Amp. We've been very impressed with it. This is big.

thorstenball's tweet image. You can now try Opus 4.5 with us in Amp.

We've been very impressed with it. This is big.


1. Present a problem/goal 2. Vector db to store I[N] context 3. Explore, Think, Execute 4. Compound their reasoning by building on previous work and reasoning. 5. Repeat.

I had to do my own.

RidgetopAI's tweet image. I had to do my own.


United States Trends

Loading...

Something went wrong.


Something went wrong.