Edward Givens

AI Tools & Growth | If it doesn’t challenge you, it won’t change you 💡 | Sharing practical AI insights, motivation & real talk | DMs open for genuine chats

[Edward Givens @sandman0302]

@bridgemindai I love 5.6 Sol so far, I don't think Ultra is working properly yet, but max and extra high are doing it. Although I do not do a lot of front end work, and I do agree with you there.

[Edward Givens @sandman0302]

@sama You guys are kicking butt out there. Keep creating the frontier!

[Edward Givens @sandman0302]

@alexabelonix I need AI to give me 3 or 4 arms too!

[Edward Givens @sandman0302]

@sflorimm It's FANTASTIC! It is on par with Fable, and sometimes better, for a fraction of the cost

[Edward Givens @sandman0302]

@OpenAI So far, I am having VERY good luck with Sol! Great work! keep the new models coming!

[Edward Givens @sandman0302]

@bridgemindai I watched you on YouTube when you were working with SOL, I was unimpressed the same as you. But then I used it on a real project, and it just knocks it out of the park for Back End work!

[Edward Givens @sandman0302]

@elonmusk Grok made a monumental upgrade. The leap is real! I am able to use Grok for pennies compared to Fable! Nice work XAI, SpaceX.

I would like to know if others have used it and what they feel, please fill out the poll :)

[Edward Givens @sandman0302]

What I’m most excited to build with GPT‑5.6 isn’t just another app—it’s a self-improving agent orchestrator that coordinates multiple AI models, learns from outcomes, and knows when to bring a human into the decision.

Better models are exciting. Better systems are the real leap.

What are you building first? 👇

[Edward Givens @sandman0302]

@acadictive I followed you :)

[Edward Givens @sandman0302]

The AI winners in 2026 won’t be the people with the best prompts.

They’ll be the people who stopped using AI as a single assistant… and started running it as a small team.

One agent for design & taste. One for complex logic & engineering. One for orchestration, decisions, and shipping.

Most people are still treating every model like a solo genius.

The ones pulling ahead are treating them like a staff.

This shift is already happening. Most just haven’t noticed yet.

#AI #AIAGENTS

[Edward Givens @sandman0302]

What I’m most excited to build with GPT‑5.6: a self-improving agent orchestrator that coordinates multiple AI models, learns from outcomes, and keeps a human in the loop when it matters.

Better models are exciting. Better systems are the real leap.

What are you building first? 👇

[Edward Givens @sandman0302]

@elonmusk As a heavy user (both Grok heavy and the actual meaning of heavy)

I see Grok as a REAL contender now. It is fast enough that having to iterate with it is no big deal. It is not always a one hit wonder, but it is so fast that I feel that it's speed makes up for having to iterate

[Edward Givens @sandman0302]

@browntechdude Let's connect :)

[Edward Givens @sandman0302]

@saqih_xi Good morning! I hope you are having a great day and are able to achieve your dreams

[Edward Givens @sandman0302]

We are all one big happy family. I love to share. Let's be a community :)

[winner @winner_xbt]

Comment HI let's follow you massively

[Edward Givens @sandman0302]

@elonmusk Grok is SO fast. I feel I have to iterate a bit more than I do with Fable, but getting something done 75% faster is ok when you have to use more than one prompt!

[Edward Givens @sandman0302]

@bridgemindai Very not bad! Imagine the next iteration Grok comes up with!

[Edward Givens @sandman0302]

I can be honest with Fable 5 and Grok 4.5, but do not have Sol yet.

I'd say the claim of Grok being as good as Opus 4.7 is probably pretty close to accurate. Maybe not at everything, but Grok is SUPER FAST. I feel people are getting used to one shot wonders. I think Grok is still something I have to iterate with, but I can still complete a task MUCH faster

[Edward Givens @sandman0302]

@elonmusk It's fast, it produces good code, and its accurate. What more can you ask for? Oh yeah, it's CHEAP! This was a huge step up

[Edward Givens @sandman0302]

What a week for frontier models 🔥

Claude (Fable 5 / Opus 4.8) Still the coding boss. Fable 5 hitting ~95% on SWE-bench Verified in recent harnesses; Opus 4.8 right behind ~88–89%. Best agentic feel, Claude Code / Cursor workflows, clean prose. Expensive but worth it when the PR has to ship.

OpenAI (GPT-5.5 / 5.6 family) All-rounder + ecosystem. Strong Terminal-Bench (low–mid 80s claims), solid science/reasoning (GPQA often ~92–93%), agent “ultra” modes rolling. Access still a bit gated on the newest tier, but daily driver for a reason.

Grok (4.5) The efficiency rocket. ~86–87% SWE-bench Verified in the same harnesses as the Claudes above, Terminal-Bench neck-and-neck, way cheaper + fewer tokens + real-time X. Feels Opus-class without the bill shock.

No single winner. Claude when the codebase is sacred. OpenAI when you need breadth + tools. Grok when speed + cost + live context matter.

Exciting week. Who are you defaulting to right now?

[Edward Givens @sandman0302]

@elonmusk It feels like I could never run out of tokens, and does what I ask. It is a completely different experience than 4.3! It does very well at taking over large tasks, where before it fell on it's face! So real agentic workflows are now possible