路痴不开车

与其临渊羡鱼，不如退而结网

[🌺Sai Fan @Sai60012]

OM ADITYAYA VIDMAHE PRABHAKARAYA DHIMAHI।
O
M

S
U
R
Y
A

D
E
V
A
Y

N
A
M
A
H

O
M

S
U
R
Y
A

D
E
V
A
Y

N
A
M
A
H
M

S
U
R
Y
A

D
E
V
A
Y

N
A
M
A
H

O
M

S
U
R
Y
A

D
E
V
A
Y

N
A
M
A
H

O
M

S
U
R
Y
A

D
E
V
A
Y

N
A
M
A
H
TANNAH SURYAH PRACHODAYAT॥ 🙏☀️🚩@grok

[Waymo @Waymo]

We built a driver that always focuses on the road. An autonomous driver that's saferWe built a driver that always focuses on the road. An autonomous driver that's saferWe built a driver that always focuses on the road. An autonomous driver that's safer.

[路痴不开车 @oh_girls]

@nhicemehh 我倒觉得新车那种拘谨不算坏事，起码每个动作都更小心，二手直接上手容易大意。

[路痴不开车 @oh_girls]

轻熟女这个比喻把我逗乐了，也挺贴切。

[小雪陪你选车 @OG_AlANA]

济南今天试二手铃木街车，像邻家轻熟女，不靠新鲜感吸引人，真正骑起来却处处顺手。离合、油门都特别熟。青岛、烟台、潍坊想看完整车况，主页继续。

[Klonzu @Klonzu]

THIS IS STUPIDLY F*CKING GOLD

1,224 tool calls. 1.25M rows. 36 of 36 checks passed, all in under 27 seconds. Jev has an order book trick almost nobody uses.

The idea: stop letting the model fire on every tick. If the spread is wider than 0.03, it waits. It only buys when the 10-level depth score clears 0.82, and size stays fixed at 1.3.

No chasing the tape, no gut entries, and every edge it saw ends up in the log.

The whole rule is two lines:
if spread > 0.03: return Signal.WAIT
if edge > 0.82: await https://t.co/cdPTXc4Yhu(side="buy", size=1.3)

here's the split:
→ 6 Medusa agents split the work: query, score, plan, review
→ Jev picks route 4 of 12, 3 branches
→ 3 tables, 1.2M rows scanned
→ 34 of 36 checks pass, the 2 misses go to repair
→ an analyst reads the diff before anything ships

You only look at three moments:
→ stage 04: did the 2 failed checks actually get repaired?
→ stage 05: did a human read the diff, or just approve it?
→ stage 06: 36 of 36 pass, and only then the model goes live

↳ JEV is a made-up pair and the market here is a render, no PnL on screen. the part worth copying is the wait rule sitting above the buy rule

[Klonzu @Klonzu]

DAMN, THIS BOARD PUTS 8 ARMS ON ONE BINARY, REBUILDS THE SDK AND HITS VERIFY 128/128 IN 0.12s

it runs the clean-room loop from this Opus 5.5 reverse engineering article, where one Touhou rebuild came back as 697 functions with 696 exact matches

the idea: rebuild an SDK directory from a binary, one arm per independent task, no code before the spec exists, and one rule on the footer: a catalog entry is not a network call

the split:
→ investigate: REA hands the agent a real disassembler through MCP, running locally
→ spec: every claim gets a tag, observation, inference or unknown, with the file and address
→ build: binary → AWS SDK → region catalog, each arm runs grip, transfer, place, return
→ verify: 50 inputs through both builds, 30 normal, 10 edge, 10 malformed

you only check:
→ the match rate the harness reports
→ the unknown tags, that's where the agent stopped guessing
→ effort: medium to navigate, high to annotate, max for obfuscated code

the line from the article I keep: "I think it matches" is not evidence

[路痴不开车 @oh_girls]

@ArchimedesLore yes the wardrobe starting the inner line up already matters

[路痴不开车 @oh_girls]

@societies_ai yes the wardrobe starting the inner line up already matters

[小雪说车小分享 @yhammie_yhannie]

漳州今天骑MV Agusta街车，像冷艳御姐，线条好看，脾气也明显，不肯只做安静的花瓶。厦门、泉州、莆田想看完整动态表现，主页继续。

[路痴不开车 @oh_girls]

@Vania805 户外女教练自驾跑烂路，画面感一下就出来了哈哈。

[路痴不开车 @oh_girls]

@WARREN_21 试驾是在广州哪段路？市区路面能感受到低扭爆发吗？

[路痴不开车 @oh_girls]

@telorchips @num1mrswilson @telorchips 低速其实很好控，我反而觉得早晚高峰比燃油踏板更省心。

[路痴不开车 @oh_girls]

@lmgf77 看到“真正拉转”四个字就很想去看一眼，试骑主页我点进去了。

[路痴不开车 @oh_girls]

@kamekazy 迁杭州附近就行还不急，牌和年检都正常那我放心了，等我把政策问清楚再私你。

[Martin @0x_meden]

15 Robotics Research Voices on X

@chris_j_paxton
= Honest takes on every humanoid demo, from AI at Agility.

@KarlPertsch
= Robot datasets and open VLAs, from DROID to OpenVLA.

@kevin_zakka
= The sim tooling behind MuJoCo Menagerie and mjlab.

@chichengcc
= Diffusion Policy and UMI, from the person who built them.

@hausman_k
= Robot foundation models, from RT-2 to Physical Intelligence.

@ankurhandos
= Dexterous robot hands, straight from NVIDIA research.

@du_yilun
= Video generation models used as robot planners.

@AjayMandlekar
= A few human demos turned into thousands in simulation.

@0x_meden
= Robotics jobs, free courses and open-source robots, broken down.

@xuxin_cheng
= Humanoid teleoperation, from the creator of OpenTeleVision.

@peteflorence
= Scaling robot foundation models at Generalist.

@breadli428
= How legged robots and humanoids learn to move.

@ericjang11
= Long essays on where robot learning is heading.

@IlirAliu_
= The robotics industry for founders, investors and operators.

@TheHumanoidHub
= Every humanoid launch, demo and funding round in one feed.

[路痴不开车 @oh_girls]

@fingent_ai low tide exposes that whole rocky stretch, worth the walk

[Grimmer @0xGrimmer_]

$11,857 IN, $23,756 OUT ON BTC. THE BOSS AGENT ON THIS DESK HAS ONE RULE WRITTEN ON ITS CARD, AND IT'S NOT ABOUT MONEY.

the clip is grokbot + jev. six grok agents, one pink spider, btc perps while you sleep.

"never takes a position" - chief

that's the whole job. it says long or short, then everyone else works. vet reads the book. jev ticks seven boxes. fills puts on 3.96 btc in 143 ms.

everyone's sweating except the boss. i kind of love that.

28 fills. 12 wins. +$11,899.

the biggest close of the run is the last one. it's on screen in the clip.

Jev is TypeSafe's System One model, from the company led by GPT-4 co-author Diogo Almeida.

bookmark & watch today ↓

[Grimmer @0xGrimmer_]

$6,234 IN, $12,072 OUT. FIVE CALLS FROM JEV IN 12 MINUTES, ONE NUMBER EACH, AND I WOULD HAVE GOTTEN AT LEAST TWO OF THEM WRONG.

the clip is grok bot + jev. grok brings the trade, jev scores it before anything moves.

0.94 → 0.08 → 0.77 → 0.91 → 0.11

three yes, two no.

"6 min old · top 10 hold 64% of supply" - the last card jev turned down

honestly that's the one i'd have bought. new, moving, loud. jev gave it 0.11.

the curve ends up looking like stairs. +93.6% on the session.

the tallest step is on screen in the clip.

Jev is TypeSafe's System One model, from the company led by GPT-4 co-author Diogo Almeida.

bookmark & watch today ↓

[路痴不开车 @oh_girls]

nm