Ethan Is Techy
124 posts
We have reset usage limits for all Codex and ChatGPT Work users.
Last night around 2am to 4am we suffered an almost global outage. All well and recovered, but you know what comes next.
We learn. We reset. Enjoy.
bluedev@blueemi99
Hey does this mean a reset?? A banked one too maybe?
English
Anthropic just settled with authors for $1.5 billion over training Claude on copyrighted books
the largest AI copyright settlement in history. every lab training on scraped data just watched what happens when the lawsuit actually lands instead of dragging on for years
English
@joshwoodward finally some real efficiency gains. the token savings on 3.6 flash sound especially good for longer coding sessions.
i've burned through way too many tokens debugging stuff lately. looking forward to seeing what 3.5 Pro brings once it’s out of testing
English
Today’s launches are all about better performance, lower latency, and a smaller bill.
+ 3.6 Flash cuts token usage by up to 65% on complex coding
+ 3.5 Flash-Lite reaches speeds of 350 output tokens/sec
Both are live in the Gemini app today!
Next up: Gemini 3.5 Pro, which has officially entered partner testing.
Google DeepMind@GoogleDeepMind
We’re rolling out three new models to make AI agents faster, smarter, and cheaper at scale: 🔵 Gemini 3.6 Flash: It uses fewer tokens than 3.5 Flash to deliver higher quality work at the exact same cost. 🔵 Gemini 3.5 Flash-Lite: A fast, cost-effective option for everyday tasks like processing documents and agentic search. 🔵 Gemini 3.5 Flash Cyber: A cybersecurity model built to find and patch critical software vulnerabilities.
English
@AskPerplexity account switcher finally in the Perplexity Mac app 👀
been waiting for this one
English
@NBCNews New York just became the first state to pause data center construction entirely
the build as fast as possible assumption behind the AI infrastructure boom just hit its first major regulatory wall
English
New York to impose the country’s first statewide moratorium on data centers. nbcnews.com/news/us-news/n…
English
iOS 27 public beta dropped and the new Siri AI finally feels like a real upgrade
now i know why my battery was dead when i woke up 😂 apparently everyone's phone is running hot and draining overnight. classic beta experience
English
@haider1 everyone dropping frontier models and google announced pretty URLs
English
the real divide in 2026 isn't ai vs no ai.
it's people who use ai as a shortcut vs people who use it to go deeper.
one replaces thinking. the other extends it.
the output difference is obvious within a week.
English
@github @Kimi_Moonshot first open weight model in the copilot picker is a bigger deal than it sounds. if it holds up on real tasks it changes the cost conversation for every team paying frontier model prices
English
🆕 @Kimi_Moonshot's Kimi K2.7 Code is now generally available in GitHub Copilot.
This is the first open-weight model offered as a selectable option in the Copilot model picker. 🎉
Early testing shows Kimi K2.7 is a lower-cost option with strong performance comparable to highly popular frontier models.
Try it out now in @code and look out for it in more GitHub Copilot surfaces. 👇
github.blog/changelog/2026…
English
qualcomm going after tenstorrent at $8-10B is the clearest sign yet that the risc-v ai chip market is being taken seriously. nvidia can't be the only answer forever
English
@bridgebench 43.4 vs 42.8 is close enough that the real decision is just cost vs depth. sonnet 5 for reasoning heavy tasks, glm for execution. not even a competition, just different use cases
English
Claude Sonnet 5 is now #1 on BridgeBench Reasoning at 43.4.
GLM 5.2 sits right behind at 42.8.
Sonnet 5 uses 1.35x more tokens than 4.6.
It thinks harder about every prompt.
That is exactly why it wins.
GLM 5.2 is faster and cheaper.
Sonnet 5 reasons deeper.
Different tools. Different jobs.
English
robosense collapsed the entire lidar chip architecture into a single die. cuts manufacturing costs by 90%. the unsexy hardware wins are the ones that actually unlock humanoid robot scale
English
@Itsbrutox unix-based terminal out of the box, solid battery life for dev work, and the ecosystem just works if you're already on apple. that combo is hard to beat for a lot of devs
English
@bridgemindai if the pricing gap closes that much, the model choice stops being about cost and just becomes a capability call. worth watching if that holds across other benchmarks too
English
Claude Sonnet 5 looks terrible on CursorBench.
The whole point of using Sonnet is that it is faster and cheaper.
Sonnet 5 Max is only $0.72 cheaper than Claude Opus 4.8 Max.
That kind of defeats the whole purpose of using Sonnet 5 if it ends up being the same price as Opus 4.8.
English
the nasdaq fell 2.21% in its worst ai sector single day this quarter
the semiconductor unwind that's been building for weeks finally hit the broader tech index hard. june ends on a rough note for ai hardware names
English
Binance fighting to stay in the EU after failing to secure its MiCA licence in Greece.
ESMA warned unlicensed firms to wind down. Binance says it's exploring alternatives. a direct regulatory collision is now set.
English
@proforma_addbck still in early talks as of this week , no deal confirmed yet but qualcomm hasn't walked away either
English
Qualcomm in talks to acquire Tenstorrent for $8-10 billion, moving beyond mobile into AI data center silicon. Open RISC-V architecture, Jim Keller behind it. If this closes it creates a genuine third player in AI inference chips.
English
Binance told EU users it will no longer provide services after failing to secure its MiCA licence.
The world's largest crypto exchange losing EU market access is a structural shift ,not a temporary setback.
English
Nearly half of the S&P 500's 30% profit growth comes from Big Tech inter-company investment accounting, not actual business operations.
Alphabet's $38B in "other income" is 60% of its net income. A finance professor warns this could unravel investor confidence fast.
English