Chauncey Hutter III

1.2K posts

Chauncey Hutter III banner
Chauncey Hutter III

Chauncey Hutter III

@ChaunceyHutter3

Hutter Software, we build apps on solana

Katılım Eylül 2020
267 Takip Edilen175 Takipçiler
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
It turns out @github has been one-shotted by LLMs and now breaks every other day. When your entire deployment pipeline is built on github and you use it multiple times a day every day for years, the most important feature of @github is it working smoothly. @github, please stop breaking because your engineers forgot how to code with their own fingers.
English
0
0
0
24
Diana Dukic
Diana Dukic@diana_dukic·
Went from scrolling 24/7 to not even wanting to log in. X just hasn’t been hitting the same lately.
English
506
154
5K
1.3M
Noah 🎈
Noah 🎈@redacted_noah·
(1/?) Introducing the Chewing Glass Explorer, a Solana explorer built for devs. I got tired of explorers, so this week I decided to vibecode my own. Sol is basically a graph database, devs want to be able to explore the full context of accounts without a million clicks.
Noah 🎈 tweet media
English
20
30
222
25K
Jacob Creech
Jacob Creech@jacobvcreech·
P-Token, SIMD-266, has been approved 🎉🎉🎉 This change will drastically reduce token transfer costs and in turn increase the capacity of the network ETA to mainnet - April
English
107
125
798
220K
SolanaFloor
SolanaFloor@SolanaFloor·
🚨BREAKING: @Solana maintained 100% uptime for 24 consecutive months without network downtime.
SolanaFloor tweet media
English
45
77
787
53.9K
dex
dex@dexhorthy·
Here’s what’s gonna happen: - you replace your code review with feedback loops (sentry, datadog, support tickets, etc) - you stop reading the code - software factory fixes everything - one day something breaks at 3am, agent can’t fix it - nobody’s read the code in 3 months - you have 3 weeks of downtime trying to re-onboard and fix it - you lose significant % of your contracts and users - your company is now dead
dex@dexhorthy

@gregpr07 this may surprise you that thus is coming from me but I think we’re in for a 1-3 year period where stuff might break at 3am and if you’re relying on loops to fix it and nobody understands what’s under the hood, you’re looking at an existential threat to your company

English
258
566
6.9K
593.6K
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
IYKYK I’m surprised it’s taken Amazon so long to figure this out.
Lukasz Olejnik@lukOlejnik

Amazon is holding a mandatory meeting about AI breaking its systems. The official framing is "part of normal business." The briefing note describes a trend of incidents with "high blast radius" caused by "Gen-AI assisted changes" for which "best practices and safeguards are not yet fully established." Translation to human language: we gave AI to engineers and things keep breaking? The response for now? Junior and mid-level engineers can no longer push AI-assisted code without a senior signing off. AWS spent 13 hours recovering after its own AI coding tool, asked to make some changes, decided instead to delete and recreate the environment (the software equivalent of fixing a leaky tap by knocking down the wall). Amazon called that an "extremely limited event" (the affected tool served customers in mainland China).

English
1
0
3
46
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
@tomjohn1028 Isn’t the cost of running the inference higher than the cost of running the transaction?
English
1
0
3
91
Tommy J ⓧ
Tommy J ⓧ@tomjohn1028·
Solana validators prioritize transactions based on fee-per-CU. But requested CU rarely matches actual CU consumed. Mean ratio of consumed/requested: 0.57. Validators are making priority decisions on numbers that are ~43% off. This is a solvable problem.
English
6
4
32
3K
Ben ⏳💧∿
Ben ⏳💧∿@HypoNyms·
One of Solana’s biggest limitations atm is asset diversity and liquidity. If someone wants to long $5M of oil futures on Solana, they can't. This changes in the next 12 months.
English
14
1
79
6.1K
Will Warren 🍵
Will Warren 🍵@willwarren·
PropAMMs have been intentionally spoofing Solana aggregators for a long time now. It's easily observed and unambiguous. But apparently it's been a taboo topic until recently? No one was willing to acknowledge it. And so nothing has been done to police it. Aggregation and meta-aggregation has created a prisoner's dilemma that rewards bad actors at the expense of the honest ones. Policing aggregator spoofing is good and necessary for our industry. It should be acknowledged and discussed. Bad actors should be named and shamed.
Will Warren 🍵 tweet media
English
10
3
51
4.2K
Nikita Bier
Nikita Bier@nikitabier·
@RaoulGMI Should be fixed once Apple approves this build. For now, you can turn your device to dark mode. Sorry about that.
Nikita Bier tweet media
English
109
32
953
185.8K
Raoul Pal
Raoul Pal@RaoulGMI·
Has anyone's else's mobile experience switched to light mode with no ability to change it back? @nikitabier
English
78
7
462
140.2K
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
@nikitabier please let me manually set a dark theme again. Please. Why are you suddenly locking the theme to system default?
English
0
0
2
21
wavey cavey ∿
wavey cavey ∿@cavemanloverboy·
ima need a CouplaBeers once we invade iran
English
4
0
5
575
941
941@level941·
Pigeon. Question. @mert if a custom solana program needs to execute two swap instructions atomically in one tx who do i call to get that done?
English
22
3
94
24.1K
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
Unleash the slop cannons
Peter Steinberger 🦞@steipete

PRs on OpenClaw are growing at an *impossible* rate. Worked all day yesterday and got like 600 commits in. It was 2700; now it's over 3100. I need AI that scans every PR and Issue and de-dupes. It should also detect which PR is the based based on various signals (so really also a deep review is needed) Ideally it should also have a vision document to mark/reject PRs that stray too far. This can't be fully automated, but even assisting would help. The closes I found is an obscure oss project. How's no startup working on this?

English
0
0
0
48
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
@aidenybai You used a Link component to prefetch /logout but the logout route/page executed serverside code to log the user out. This is not a NextJS issue.
English
0
0
4
1.7K
Aiden Bai
Aiden Bai@aidenybai·
we had this insane P0 bug users would get randomly logged out we had a bunch of users churn... ...turns out Next.js prefetching the /logout page
English
91
24
1.8K
605.1K
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
@bqbrady Use a multi-core build server with a local build cache on the server. I do this and my rust builds are extremely fast.
English
0
0
0
35
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
@RobotsWon @AnthropicAI We’re talking about knowledge and reasoning and you’re bringing up physical capabilities. Boston dynamics already made a robot that can do a backflip. That’s not what we’re taking about.
English
0
0
0
62
Anthropic
Anthropic@AnthropicAI·
New Engineering blog: We tasked Opus 4.6 using agent teams to build a C compiler. Then we (mostly) walked away. Two weeks later, it worked on the Linux kernel. Here's what it taught us about the future of autonomous software development. Read more: anthropic.com/engineering/bu…
English
873
2.5K
21.4K
8.5M
Chauncey Hutter III
Chauncey Hutter III@ChaunceyHutter3·
To be honest. I think that it emulated reasoning in a way. But the architecture is the same. Our brains reason, circuits can reason if we encode logic gates into them, but an LLM is still just generating tokens. It can emulate correct reasoning with enough tries and if we tell it enough logic rules that it can pattern recognize. Which is clearly enough to solve the above problem. I am optimistic that researchers will discover a new architecture that is more capable of true reasoning, but imho I don’t think LLMs in their current form can get us there.
English
0
0
0
50
vibebuilder
vibebuilder@vibebuild·
@ChaunceyHutter3 @SeShpk @AnthropicAI Yes you are wrong that they don't understand or reason. I would argue that a LANGUAGE model being able to solve novel Erdos problems is a very clear indicator that the model is clearly understanding and reasoning, would you agree? If you disagree, explain.
English
1
0
0
50

Keşfet

@github @elonmusk @diana_dukic @redacted_noah @jacobvcreech @SolanaFloor @solana @dexhorthy