Ender

244 posts

Ender banner
Ender

Ender

@ender_mode_

Yes, it’s me.

San Francisco, CA Katılım Eylül 2025
152 Takip Edilen40 Takipçiler
Ender
Ender@ender_mode_·
@claudeai I will eat a hat on live camera if you show a video of it operating a computer any where near the speed and accuracy in that demo video
English
0
0
0
427
Claude
Claude@claudeai·
Available on Pro and Max. Update your desktop app and pair with mobile to try: #dispatch-and-computer-use" target="_blank" rel="nofollow noopener">claude.com/product/cowork…
English
55
55
1.8K
553K
Claude
Claude@claudeai·
You can now enable Claude to use your computer to complete tasks. It opens your apps, navigates your browser, fills in spreadsheets—anything you'd do sitting at your desk. Research preview in Claude Cowork and Claude Code, macOS only.
English
3.4K
9.4K
93.9K
34.4M
Ender
Ender@ender_mode_·
Have we just accepted that all AI demos are *at best* speed up tremendously and at worst downright fruad
English
0
0
0
4
Ender
Ender@ender_mode_·
@felixrieseberg Hey Felix, love the product but don't you think it's a bit disingenuous to show a demo video that implies a speed of computer use we both know Claude is no where near achieving. Can you publish how long those tasks would actually take?
English
0
0
1
176
Felix Rieseberg
Felix Rieseberg@felixrieseberg·
Today, we’re releasing a feature that allows Claude to control your computer: Mouse, keyboard, and screen, giving it the ability to use any app. I believe this is especially useful if used with Dispatch, which allows you to remotely control Claude on your computer while you’re away.
English
754
1.2K
16.2K
3.5M
Ender
Ender@ender_mode_·
@TaylorOgan I mean there is an absurd amount of evidence this is just false - At least for the majority of his children. But you do you man
English
0
0
0
906
Ender
Ender@ender_mode_·
It’s truly amazing how selling a nearly commoditized service (compliance) brings out the worst in founders (fraud)
English
0
0
0
11
Ender
Ender@ender_mode_·
@BryanOnel86 Crazy it took this long, hilariously well known in SF startup community that this was basically Deleve’s main selling point to founders who just wanted to slap the SOC2 logo on the website
English
0
0
17
2.3K
Bryan Onel
Bryan Onel@BryanOnel86·
Yeah this doesn’t surprise me in any way, given the history I have with the founder of Delve. What a bombshell of an article though. There is just so much to unpack.
erin griffith@eringriffith

A detailed and brutal look at the tactics of buzzy AI compliance startup Delve "Delve built a machine designed to make clients complicit without their knowledge, to manufacture plausible deniability while producing exactly the opposite." substack.com/home/post/p-19…

English
20
10
561
67.1K
Ender
Ender@ender_mode_·
People are merging > 50% of the code GPT 4.1 outputs… All these outages are starting to make sense
Kyle Daigle@kdaigle

Hot take from looking at @github Copilot telemetry: benchmarks make coding models look wildly different. Production workflows make them look much more similar. 👀 We looked at 23M+ Copilot requests and examined one simple metric: code survivability.

English
0
0
0
38
Ender
Ender@ender_mode_·
@kdaigle This just tells me the amount of bugs being committed to repos is skyrocketing
English
0
0
0
7
Kyle Daigle
Kyle Daigle@kdaigle·
The team’s been cooking up additional insights here. More to come if this is interesting to folks. 👀
English
6
0
9
2K
Kyle Daigle
Kyle Daigle@kdaigle·
Hot take from looking at @github Copilot telemetry: benchmarks make coding models look wildly different. Production workflows make them look much more similar. 👀 We looked at 23M+ Copilot requests and examined one simple metric: code survivability.
Kyle Daigle tweet media
English
28
49
335
73.8K
Ender
Ender@ender_mode_·
@garrytan It is actually amazing to watch in realtime the head of YC speed run every thing that YC says not to do
English
2
0
65
4.6K
Garry Tan
Garry Tan@garrytan·
This new release of GStack is for all the haters on Product Hunt who said it was just a bunch of markdown files
Garry Tan tweet media
English
108
40
810
105K
Ender
Ender@ender_mode_·
@garrytan This must be a bit right?
English
0
0
1
94
Garry Tan
Garry Tan@garrytan·
My CTO friend texted me: "Your gstack is crazy. This is like god mode. Your eng review discovered a subtle cross site scripting attack that I don't even think my team is aware of. I will make a bet that over 90% of new repos from today forward will use gstack."
Garry Tan@garrytan

gstack is available now at github.com/garrytan/gstack Open source, MIT license, let me know if it works for you. It's just one paste to install it on your local Claude Code, and it's a 2nd one to install it in your repo for your teammates.

English
146
50
1.7K
869.3K
Chris Tate
Chris Tate@ctatedev·
Credential brokering is the most underrated feature of Vercel Sandbox The sandbox never receives the secrets, eliminating an entire class of exfiltration risks I've had to build this pattern over and over for sandboxed agents and untrusted code Now just a few lines of code
Chris Tate tweet media
English
26
17
289
27.6K
Ender
Ender@ender_mode_·
Its ironic that it seems AI coding agents are the only ones not aware of how much less time it takes to build things
English
0
0
0
52
Ender
Ender@ender_mode_·
@BrendanFalk Could include search tool too depending on observed accuracy of above or dynamically based on semantic similarity scores. Again, tbh I’d just chuck them all in the prompt and call it a day in vast majority of cases
English
0
0
0
13
Ender
Ender@ender_mode_·
@BrendanFalk Embed all skill descriptions before and then at runtime do similarity matching for most likely needed skills. Some nuance here depending on expected prompt lengths / complexity so it may make sense to chunk a bit. Then just include top skills + semantic matches
English
1
0
0
9
Brendan Falk
Brendan Falk@BrendanFalk·
Question for AI engineering community: what is the current best practice for giving a single agent access to a potentially unbounded number of skills? Goals are (in priority order) 1. Maximize skill use accuracy 2. Minimize context use 3. Minimize unnecessary tool calls
English
36
4
88
56.3K
Ender
Ender@ender_mode_·
@garrytan @garryslist That’s awesome! Wasn’t trying to hate on the project just felt like in an AI coding agent world, LOC is largely meaningless
English
1
0
4
2.7K
Ender
Ender@ender_mode_·
@mlejva @opencode @modal works better. Your start up times were good but intermittent connectivity, restoring sandboxes, and long running sessions may have had one 9 of consistency
English
2
0
2
338
Vasek Mlejnsky
Vasek Mlejnsky@mlejva·
Looking to run @opencode in a sandbox? We have a ready-made guide and template $> 𝚎𝟸𝚋 𝚜𝚋𝚡 𝚌𝚛𝚎𝚊𝚝𝚎 𝚘𝚙𝚎𝚗𝚌𝚘𝚍𝚎
Vasek Mlejnsky tweet media
English
7
8
188
13.4K
Ender
Ender@ender_mode_·
@deedydas You either don’t understand what Azure does or what OpenRouter does but this is a comically apple to oranges comparison
English
0
0
0
35
Deedy
Deedy@deedydas·
OpenRouter weekly token consumption has exploded 12.7x in the last year to ~12.1 trillion tokens/week or 662T tokens/year run rate. To give you a sense of this scale, Azure did 100T tok/quarter 6mos ago. OpenRouter is responsible for about as much inference as ALL of Azure.
Deedy tweet media
English
68
84
868
225.6K