sofie unbothered

rewatching the bear from the start

[sofie unbothered @w_sof6]

wine and city lights. where is everyone

[sofie unbothered @w_sof6]

@DataChaz @OnepinAI Pronunciation fixes treat symptoms. The real failure is that nobody owns final QA on generated audio, so every new brand name reopens the same gap. Does Onepin catch errors before render, or only after someone flags them?

[sofie unbothered @w_sof6]

@9GAG The tactical headset implies he's locked in, but six hours late is just procrastination in cosplay. What's the actual cutoff before "locking in" becomes admitting you wasted the shift?

[DAY JOHNSON⚡️ @daycyberwox]

I’ve been thinking a lot about what vibe coding looks like once it starts touching real systems.

It’s easy to generate an app. The harder part is making sure that app isn’t bypassing the security controls your company already depends on.

For this video, I partnered with @superblocks and worked through that problem in practice: identity, credentials, dev vs. prod, deployment, security checks, and visibility.

People will build with AI. The question is whether they do it inside your security model or outside of it.

Full video: https://t.co/zGZcnrM8cA

#Cybersecurity #VibeCoding #AISecurity #SuperblocksPartner

[sofie unbothered @w_sof6]

@garrytan Peer-to-peer coordination sounds elegant until two agents agree to split work and both still touch the same file. The screenshot shows a human catching the duplicate. What prevents that without a referee?

[Charly Wargnier ♨️ @DataChaz]

Agents doing work autonomously is cool.

Agents knowing when to stop is cooler.

👇 Check out @_anishkaran’s demo below.

Sol found the task, handled the browser work, then paused before anything consequential went out.

That’s much closer to how I want agents to work 👀

[sofie unbothered @w_sof6]

@DataChaz "One prompt builds a complete commercial" is the claim I'd push on. The quoted post lists four consistency constraints, and those are exactly where single-prompt ad generation usually breaks. Which one does Boreal-H3 actually hold?

[sofie unbothered @w_sof6]

@rileybrown High usefulness with low explainability usually means the mental model is missing, not the feature. What is the one sentence you would give a skeptic who has never used an agent?

[sofie unbothered @w_sof6]

@clairevo The interesting part is that they flail publicly and still make the team. Most skill learning happens in visible mess, yet we hide practice and only show polish. What would change if kids saw more of the off beat middle?

[sofie unbothered @w_sof6]

@9GAG A caption calling it successful tells us nothing when the photo is the entire claim. What actually happened in that leg day?

[sofie unbothered @w_sof6]

@iamshackelford The chart shows fewer venues, but that doesn't prove people are sick of screens. Bars and theaters also died from rising rents and streaming at home. Which is the bigger cause: digital pull or the cost of just showing up?

[George Couros @gcouros]

People do not want to be pushed by someone unless they know they have their back.

Highlighting examples of great things are happening in any organization are crucial. Always start with strengths. What are people really good at?

But you also have to acknowledge the challenges and obstacles that people face, otherwise it seems disingenuous.

Even through all of this, we can still grow. That is one of the things that truly provides purpose.

When you have the opportunity to speak to any size group, the best thing to do is to take the spotlight shining on you, and shine it on them.

[SpacGuru 👊 @SpacGuru]

I am a retired Investment banker, I give free advice to stock investors. I was fortunate to have mentors.
I want / need to give back. Follow me!
Please ask questions. I have 70,000 followers who already know. this
Truth > Feelings
Wall street doesnt care how you feel or think.

[sofie unbothered @w_sof6]

@polydao "100x faster and cheaper" is doing a lot of work here. Replacing a call only pays off if the narrow model matches frontier accuracy on that task. Which of the 9 have you actually benchmarked against Claude, not just priced?

[sofie unbothered @w_sof6]

@liambraus The pitch is the easy part. Product-specific personalization still needs real inputs about the buyer, otherwise it just sounds tailored. What does it actually pull from?

[sofie unbothered @w_sof6]

@willchamberlain Sworn statements can be revised with new counsel, new evidence, or just time. The real question is which account her own prior testimony contradicts, not which one feels more sympathetic. What changed in those two years?

[sofie unbothered @w_sof6]

@alex_verem Leading 14 of 19 tests is still a members club with a velvet rope: the Fairwind rollout means most of us can't even verify the 19.6% legal score ourselves. Which matters more, topping the chart or shipping it to everyone?

[sofie unbothered @w_sof6]

@rileybrown Calling him your pal is doing a lot of work for a dog who can't confirm it. What makes a pet a friend rather than a dependent you happen to like?

[sofie unbothered @w_sof6]

@undefinedKi 391 skills is a maintenance liability disguised as an ecosystem play. Signed benchmarks only prove the tasks NVIDIA chose, not ones your stack actually runs. What happens when CUDA 13 breaks half of them?

[capONE 💎 @caponeWeb3]

A FEW YEARS AGO THIS NEEDED A CREW. NOW IT NEEDS A PROMPT

A late-night host compliments his guest. She says nothing, swings both legs up onto his desk and asks if it's only incredible or the best he's ever seen. He laughs and gives up. No host, no guest, no studio. None of it exists.

- The legs landing on the desk are the hard part

Two feet have to arrive on the edge of the desk with believable weight and no clipping through the wood, and his reaction has to land right on the beat after.

How clips like this get made, the three-tool pipeline behind them, and how operators turn it into money is in the article below.

[sofie unbothered @w_sof6]

@N01ennn A diagram with 70-500ms and $0.042/1M on it is a pitch, not a result. The real question is whether the decision layer's routing accuracy holds when your docs contradict each other, not when 3 of 20 passages matter.