Post

@python_xxt 根本不是这样。我研究机器学习的,opus拉完了。Claude只适合做前端,稍微复杂一点的逻辑分析就开始hallucinate。
中文

@python_xxt AI时代了,作为熟练使用AI模型甚至用来工作的人,稠密与混合专家模型的区别这种基础的基础要了解,不了解实在是不应该
不是说要了解到多深,就像数学,不一定掌握高数,但加减乘除这种基本的运算和规则还是应该了解
中文

@python_xxt 太聪明就开始瞎搞啊,我们用模型就像雇人一样,不是越聪明越好,最关键的是要能理解我的意图,并且忠实的执行,太聪明了,他会觉得他比我厉害,然后就会歪去我的意图按照他的方向做。这跟用人也是有些相似之处的。
中文

@python_xxt 有多深不清楚,不过以Opus如今拉跨的会意能力,我凭什么给它机会聊。
但就聊天来说,Opus已经逐渐Gemini。喜欢武断,不喜欢搜索,而是在措辞上让你觉得它查过了,或者它懂,直到你指着它鼻子骂。
Codex但凡看到自己语料库没有的内容,第一时间搜
简单说,Opus拉完了。
中文

@python_xxt 那可以试试我最近用的这个项目,多个agent编排在一起,还是和终端里一样cb6c994a.hive-site.pages.dev
跑在浏览器里
中文

$NXXT exploded +134% in just 55 minutes while volume surged past 51.58 million shares traded versus a 65-day average of only 3.14 million.
The move was not random. Q1 2026 revenue climbed to $21.1 million from $16.3 million, up 29% year over year, while gross profit jumped to $1.71 million from $518 thousand.
Gross margins also expanded sharply from 3.2% to 8.1%, showing the company kept significantly more money from every dollar of sales. The chart showed a vertical breakout followed by consolidation instead of an immediate collapse, which traders watch closely in momentum names.
This is why small-cap stocks with rising revenue, improving margins, and extreme relative volume can reprice incredibly fast in a single session. Now traders are watching the next market open closely to see whether momentum and volume continue or fade after one of the wildest moves of the day.
$NXXT $NVDA $MSFT $AMD $SMCI $PLTR $TSLA $META $AMZN $GOOGL $AAPL $FCX $HBM $SCCO $TECK $RIO $BHP $NRED $SMR $OKLO $CCJ $LEU $MSTR $COIN $BTC $ETH $SPY $QQQ $DIA $IWM $SOUN $BBAI $IONQ $RKLB $ASTS $HOOD
English


















