Sai.Eth⛩

Into Web3 & crypto
Focused on learning, building & growth.

[Sai.Eth⛩ @Sai_ethym]

Tonight I learned that every improvement starts with noticing a problem.
The biggest takeaway for me was to look at everyday challenges differently and ask, “What could be better?” 🔓

#UNLOCK

[Sai.Eth⛩ @Sai_ethym]

@inference_labs Turning predictions into actual actions is key.

[Inference Labs @inference_labs]

Inference is only useful when something happens next.

Sertn workflows can route model outputs into actions, Slack, email, or other downstream systems, turning a prediction into an operational response.

[Sai.Eth⛩ @Sai_ethym]

@inference_labs That's a wild thought about inference costs.

[Inference Labs @inference_labs]

Model prices can fall while AI bills still rise.

Gartner expects inference cost per agentic workflow to grow more than 5x as agents reason, retry, call tools, and run continuously.

The new unit of AI economics is not the model subscription. It is the inference.

[Sai.Eth⛩ @Sai_ethym]

@inference_labs Wow, that's a concerning development.

[Inference Labs @inference_labs]

AI is now finding cybersecurity weaknesses faster than some financial firms can remediate them.

That flips the bottleneck.

Detection becomes abundant. Prioritization, evidence, and deciding which machine-generated findings deserve action become scarce.

[Sai.Eth⛩ @Sai_ethym]

Been using USDD to pay gas in @TokenPocket_TP countless times, and the experience has been smooth and convenient. No need to keep extra tokens just for gas.

@USDDecentralize @KeyPalWallet
#USDD #TokenPocket #GasBoostSeason #USDDonTP #USDDForGas

[Sai.Eth⛩ @Sai_ethym]

@UtribeOne That's a really good way to put it.

[dkarbon @dkarbon_green]

You buy it to destroy it.

That is the only moment a carbon credit does what it exists for. 168 million were retired last year, and each one can never be sold or claimed again.

How it works 👇

#dkarbon

[Sai.Eth⛩ @Sai_ethym]

@inference_labs Interesting take on that.

[Inference Labs @inference_labs]

2/ That opens up more specific workflows than simple detection.

Was the correct signal given? Did the gesture change? Was the worker positioned safely relative to the aircraft?

Sertn can turn those visual movements into structured model outputs.

[Inference Labs @inference_labs]

1/ Keypoints are useful when the important signal is not just who or what is in the frame, but how something is positioned and moving.

For airside operations, that can mean tracking a marshaller’s arms, body position, and gesture sequence over time.

[Sai.Eth⛩ @Sai_ethym]

@Timthadon01 Sounds like a major step forward.

[Bahdfella™ ❤️‍🔥 @Timthadon01]

This is a big deal for Physical AI.

We’ve spent so much time talking about what models can see, but hearing the real world is a completely different challenge.

BRIDGE ASR 2.0 is testing 23 models across 23 languages and real conversation conditions like interruptions, overlapping speakers and code-switching.

That’s the kind of evaluation that matters when AI has to work outside a perfect lab environment.

Curious to see where the models actually stand. 👀

[Sai.Eth⛩ @Sai_ethym]

@UtribeOne Agreed, teachers make all the difference.

[Sai.Eth⛩ @Sai_ethym]

@UtribeOne Finance doesn't need to be scary.

[Sai.Eth⛩ @Sai_ethym]

AI safety needs proof, not promises.

@SertnAI uses proof of inference to verify what AI models actually did.

As AI gets smarter, trust alone isn’t enough.

#VerifiableAI #PacingTheFrontier #ProofOfInference #AISafety #AI

[Sai.Eth⛩ @Sai_ethym]

@inference_labs Makes sense. So the *real* skill is choosing good answers.

[Inference Labs @inference_labs]

Mathematicians are starting to talk about an era of “proof abundance.” That is an interesting preview of AI more broadly.

When producing answers becomes cheap, the scarce thing is no longer output. It is knowing which outputs deserve to be accepted.

[Sai.Eth⛩ @Sai_ethym]

@UtribeOne This is a tough one, but I'd lean towards supporting growth. A growing economy generally means more opportunities for everyone.