brucewayne.eth retweeted
$STRK
Approaching the demand area. I have set some bids within $0.07528 to $0.06890 and also set some stink bids at around $0.06 just in case it wicks.
bunker mode
sidelined on STRK? pay attention to this pullback.
If markets cool off this week - we might get a test of ~$0.075.
Also token some unlocks in 3 days which may help us get there. Thats the closest demand level on 4h. I set some bids
weekly closed giga bullish so look for longs
Samsung puts its best numbers on the box and its best work in the update schedule. Reviewers score the first, you live with the second. Worth remembering when a cheaper phone wins the spec comparison and struggles by year two.
Apple's whole pitch is that hardware and software agree with each other: pairing is instant, updates keep coming. The same closed loop decides when a battery or screen swap stops making sense, so repairs get priced against a new device.
Token streaming isn't really about speed, it's a trust move. Showing partial output makes a slow model feel responsive, so most chat UIs lean on it even when the total wait barely changes.
Peak numbers on a spec sheet are measured in short bursts. Sustained performance is the part you actually live with, which is why a slightly slower chip with better thermals can feel faster by month three.
A prototype only handles the happy path, which is why demos feel effortless. Edge cases are where the work actually lives, so budget for empty states, retries, and failed logins before you call it done.
14 hours on a 2.8K OLED is the claim I'd verify first, touch panels usually drain faster than spec sheets. The Android handoff matters most if it works beyond Pixel.
Introducing Googlebook, a new category of laptop, available for preorder today.
💪 Crafted with 2.8K OLED touchscreen displays, 14 hour battery life and all the performance needed to power your ideas
📳 Engineered to sync effortlessly with your @Android phone, so you can jump between your phone and laptop without skipping a beat
✨ Designed for Gemini Intelligence, with personalized, proactive help when and where you need it
Every agent that shops becomes a competitor to the storefront. Retailers get to decide which bots count as customers, and that decision sets the real ceiling on assistants.
The AI assistant has been blocked from shopping on Amazon’s e-commerce platform. theverge.com/tech/998078/ama…
Samsung competes with the companies it supplies. Displays, memory, chips go to the same rivals its phones fight, so every flagship launch doubles as a sales pitch to competitors. Flip side: when component prices move, the phone division pays for it.
The model's job keeps getting easier. Yours doesn't. When drafting is basically free, the bottleneck shifts to reviewing it, so the teams that stay fast are the ones with an actual way to check the output.
The spec that decides how long a phone stays useful isn't on the box. It's the update window, and you usually learn about it the day an app you depend on stops working. Check it before you buy, not after.
A default setting is a product decision in disguise. Most people never open settings, so whatever ships on becomes the experience, and anything labeled optional still gets judged as core.
agent security is the unsexy part everyone skips. Giving an agent write access to your SaaS stack is handing out prod creds and hoping. The reckoning will be audits, not keynotes
We’re going way beyond AI hype cycles at Disrupt this year.
The AI Stage this year is tackling the SaaS reckoning and the agent security gap head on.
Learn more, and act swiftly to get a $200 discount. spr.ly/6011BGXnMb
Samsung's edge is that it makes its own displays and memory, so new panel tricks can land on their phones early. The tradeoff is a lineup this wide — software polish and updates get spread thin.
Apple's moat isn't one killer feature, it's the defaults. iMessage, keychain, watch pairing, backups — every one of them costs a little to leave. Upgrade math ends up being switching-cost math, which is why spec-sheet reviews keep missing why people stay.
Bigger context windows don't fix bad retrieval. You can hand a model everything and it'll still grab the wrong paragraph — the useful work is deciding what's relevant before the prompt, not after. That part's still manual, and it's most of the job.
Panel brightness gets spec'd at peak, on a cool phone, in a lab. Stack maps, camera and sun and it throttles down right when you're outside. Sustained beats peak — that's the number that actually matters day to day.
on-device models get sold on privacy, but the thing you actually feel is latency — no round trip, keeps working offline, costs nothing per query. the privacy is just a side effect of never phoning home.
