Tweet épinglé
Sami | SEO AI
15.2K posts

Sami | SEO AI
@samigrows
Helping businesses get quality traffic with SEO | Worked with multiple international clients l DM for SEO services |
🟩🟩🟩🟩⬜⬜⬜⬜⬜⬜ Inscrit le Eylül 2024
479 Abonnements521 Abonnés

OpenAI just dropped GPT 5.6 benchmarks.
GPT 5.6 Sol Ultra hits 91.9% on TerminalBench 2.1.
That mogs Mythos 5 at 88.0% and Fable 5 at 84.3%.
Even base Sol (88.8%) beats both Anthropic frontier models.
The unlock is a new ultra mode that uses subagents to accelerate complex work.
OpenAI quietly built multi-agent orchestration directly into the model.
Pricing: $5 input / $30 output per million for Sol.
Limited preview only.
Government gated rollout customer by customer.
The frontier just moved.
The US just locked it behind a velvet rope.

English

@samigrows But if you’re paying for SEO, ideally you don’t invest when you’re bottlenecked. Fix your review velocity, THEN advanced SEO
English

I told a business this week they shouldn't hire us for SEO yet.
They had 3 reviews.
Great company, good work, decent website. They wanted to invest in ranking higher, and they had the budget for it.
My honest answer was not yet. Because even if I got every page ranking number one, it wouldn't matter. Two companies sitting side by side in search results, one has 3 reviews, the other has 150. Nobody calls the 3. We'd be getting paid to send them traffic that bounces straight to a competitor with more social proof. Rankings without reviews is a bucket with a hole in it.
Build the reviews first. Then ranking actually turns into phone calls.
Here's the part most owners miss about why reviews matter beyond just looking trustworthy. Reviews are one of the biggest factors in whether you land a spot in the map pack, which is where the majority of local service calls actually come from. Your total count, your rating, and especially how fast new reviews are coming in all determine whether Google gives you one of those three spots.
Reviews pull double duty. They get you into the map pack, AND they get people to actually call once you're there.
If your profile has been sitting at the same review count for months, that's not something to clean up later. It's the thing holding everything else back. Fix that first, then invest in ranking.

English

Welcome a new member of the 𝕏 family.
@XMoney is starting its rollout to a select group of US Premium+ members as we refine and elevate the experience ahead of a wider launch.
Excited to see it out in the wild.
Feedback is crucial - let’s make it better.

English

@samigrows I agree with you, and I'd also add the Core web vital related errors like slow loading speed to the 403/ 500 error.
Based on my SEO experience I've seen that fixing the website loading solves 90% of the problems
English

#4: Backlinks:
We focus on getting relevant backlinks instead of AI automated slop that does nothing.
English

this is funny as hell.
Elon just admitted there's too much "Starsh*t" at SpaceX.
bro named everything after the prefix "star" and finally cracked on the last one. Starship, Starlink, Starbase, Starfactory, Starshield, Starfall, Stargaze, Starmind. he basically built a whole empire off one word lmao.
the rocket scientist finally cringed by his own naming convention. 💀

Elon Musk@elonmusk
@SawyerMerritt We’re cutting back on the use of “Star” as a prefix. It’s getting a bit silly. Too much Starshit!
English

@bridgemindai I don't care who wins I just want to make a SAAS which does $5k MRR
English

Open source AI needs to catch up. Now.
Claude Fable 5 scores 76.5 on the Artificial Analysis Coding Index.
GLM 5.2, the best open weights model in the world, scores 68.8.
That gap is closing fast.
Within six months an open source model will match Fable 5.
The trajectory is undeniable.
Meanwhile the US government is suspending American frontier models and gating GPT 5.6 customer by customer.
China is releasing weights to anyone with a download link.
If this continues, open source Chinese models will catch the US frontier just as the US locks itself out of its own.
The US needs to stop regulating.
Or China wins by default.

English

GLM 5.2 is good. It is also very far from Fable 5.
Open source will catch up within the next 6 months.
The trajectory is undeniable.
But right now, tests like the BridgeBench lava lamp make it obvious that something is still missing from open source models.
Fable 5 produces output that is more reliable, more efficient, and meaningfully more polished than GLM 5.2.
The gap is real today.
It will not be real in six months.
English

@samigrows Yes, it's definitely a special moment! Now I know I mustn't give up :)
English

🥳FIIIIIIIRST SAAAAAAAAAAAALE! 🥳
We finally did it!
The notification we’ve been waiting for just popped up, it feels incredible after all the hard work.
A few weeks ago, we took a leap and pivoted our entire offer.
Today, we have our first real proof that it was the right move.
To our first client:
We’re going to treat you like royalty. Your project is amazing, and we’re going all out to over-deliver.
Motivation is at 1000%. (And this is just the beginning!)

English

Month 1: wrote content nobody read.
Month 2: wrote content nobody shared.
Month 3: wrote content nobody linked to.
Month 4: Google decided I was the authority.
Trust the process before the process trusts you.
Sami | SEO AI@samigrows
Your website looks professional. Your Google Search Console looks empty. Design gets you compliments. SEO gets you customers. You need both but only one pays the bills.
English







