I comment on tech with brutal honesty ex-@Huawei Chips architectures @zortex in @tsinghuaxlab https://xitter.cf/t.co/6av4rzTZWW

Shibuya-ku, Tokyo
Joined April 2022
Did Deepseek just amplify the pricing war? Deepseek V4.1 flash apparently codes at Opus 5 level at 2% of the cost. Anthropic has already made Opus 5.5 dirt cheap in its subscriptions. OpenAI is serving near unlimited Sol 6.1 in their subscriptions. Now we have deepseek serving a 200tps+ model basically for free. These labs are racing to the bottom.
74
26
5
645
33,735
\横浜優勝/ \宮﨑優勝/ \成瀬優勝/ \三森優勝/ \井上優勝/ \林優勝/ \レイノルズCS5S/ クライマックス4連勝⭐️⭐️⭐️⭐️ 日本シリーズ行けるんだああああああ😭😭 皆で掴んだ勝利だ✌️‼️ そしてセ王者🐯やはり強かった❗️👏 来年もヨロシク甲子園🫡 ホークス倒せベイスターズ💪‼️🔥 #baystars
2
66
1,211
2026 indiehacker playbook: > complain on X about how [entrenched SaaS product] is so expensive now > people like and agree > idea validated! > build 83% cheaper competitor To be fair, this is a pretty decent playbook.
89
10
4
266
22,368
🟦2026/27明治安田J1リーグ 第10節🟦 ⏰70分 選手交代🔁 IN:24番 #山﨑希一 OUT:39番 #山本隼大 🆚#浦和レッズ 🎥#DAZN 視聴 ➡tidd.ly/3RKBU0f #1017浦和戦 #水戸ホーリーホック #インクデザイン
12
2
80
8,512
kun retweeted
97. Down to the final strike of a potential game-ending out in the top of the 9th, trailing by one with Japan Series on the line. Mimori’s timely single flips it to a lead and sparks a 4-run explosion. Pure clutch chaos.
1
18
🚨 ChatGPT can now access your Google Chrome tabs to complete tasks.
48
135
8
4,362
70,759
AGI baby - captchas and pizzas 🤣
Replying to @adcock_brett
hark can fill out captchas and get you free pizza, codex won’t do either, thanks for pizza
30
5
2
193
29,254
Mikel Brown Jr. in his preseason debut: 15 PTS 4-7 FG 3 REB 2 STL 2 3PM
18
144
31
1,146
47,925
Happy Roadster unveil day to those who celebrate
55
86
10
1,856
63,552
"We treat every game the same ... it's just our identity, we're going to fight until the 27th out." Garrett Mitchell joins SportsCenter with @notthefakeSVP after his solo HR gave the Brewers a 4-2 lead in the 7th in their win against the Dodgers 😤
49
29
31
560
254,261
I was genuinely hopeful for Gemini 4 Argon. There are some good parts, but generally speaking, this model should be avoided.
Gemini 4 Argon has proven to be outright unusable for real code work in my codebases. It aggressively hallucinates about how systems work.
79
23
25
966
140,222
So cool seeing our weird UX experiments in T3 Code getting validated by the labs
We've been working hard on making the Claude Code desktop app great for running lots of sessions side by side. It's now much quicker to start sessions, easier to split work across a few, and you can always see what each of them is doing. Here’s what's new:
98
31
5
2,964
135,789