erel vanono

29 posts

erel vanono

erel vanono

@ErelVanono

Backend engineer

Katılım Haziran 2019
85 Takip Edilen3 Takipçiler
erel vanono
erel vanono@ErelVanono·
@gill_kyle Totally agree, but if this is the case don’t you think that TUI solutions like Claude code would be better in that case ?
English
1
0
0
59
Kyle Gill
Kyle Gill@gill_kyle·
overall I guess probably the right call, I don't use the Cursor IDE hardly anymore and the apps that manage a lot of parallelism are probably increasingly where people are moving
English
1
0
37
11.3K
erel vanono
erel vanono@ErelVanono·
@vivoplt The only advantage I see for cursor is its autocomplete. But other then that Claude code is main tool for my work
English
0
0
0
929
Vivo
Vivo@vivoplt·
be honest, which AI tool is best for coding?
Vivo tweet mediaVivo tweet mediaVivo tweet mediaVivo tweet media
English
150
6
397
51.8K
erel vanono
erel vanono@ErelVanono·
Is cursor 3 worth the hype to come back from Claude code ? Sorry but I am thinking that my workflow is so dependent on git worktrees and using heavy gui seems to me to slow
English
0
0
0
50
erel vanono
erel vanono@ErelVanono·
Why “agent” mode is decided to be the way to go for modern dev experience ? It seems like that only cursor deliver value across the different interfaces devs using solution where beside generating code devs also editing and reviewing code Other solutions that you can suggest?
English
0
0
0
17
erel vanono
erel vanono@ErelVanono·
@ThePrimeagen Totally agree. I don’t understand how there aren’t alternatives product to supermaven that will be agnostic to different delevoper tools (vscode, nvim etc.) and different AI models
English
1
0
0
80
ThePrimeagen
ThePrimeagen@ThePrimeagen·
i am using supermaven again and i have something to say about this whole AI thing. I think as a group (swe) we rushed so fast into Agents when inline autocomplete + actual skills is crazy. A good autocomplete that is fast like supermaven actually makes marked proficiency gains, while saving me from cognitive debt that comes from agents. With agents you reach a point where you must fully rely on their output and your grip on the codebase slips. Its insane how good cursor Tab is. Seriously, I think we had something that genuinely makes improvement to ones code ability (if you have it). Truly acts as a multiplier, and we left it in the dust because it is not sexy. hurts me on the inside.
English
217
133
3.7K
183.5K
erel vanono
erel vanono@ErelVanono·
Is there anything close to the experience of cursor autocomplete in nvim ? This is the only reason that I moved to cursor And I missing so much for my nvim setup
English
1
0
2
58
erel vanono
erel vanono@ErelVanono·
@mattpocockuk Such a good take, I really believe that right quality PRD’s and manage documentation are key to thrive in SW in the AI era
English
0
0
0
23
Matt Pocock
Matt Pocock@mattpocockuk·
At this point I'd buy a $500 course on writing great PRD's
English
36
2
168
27.9K
erel vanono
erel vanono@ErelVanono·
@ForrestPKnight The real lazy ones don’t use Lazygit they just using git aliases like git status = gst
English
0
0
0
12
Forrest Knight
Forrest Knight@ForrestPKnight·
Y'all keep saying Lazygit this, Lazygit that. But what’s lazier than just typing the command in the terminal?
English
5
0
21
1.7K
erel vanono
erel vanono@ErelVanono·
Is Claude’s coding capability worth it on the Pro plan, or are the real benefits only noticeable with the Max plan and Claude Opus? I tried the Pro plan with Claude 4 Sonnet, but it didn’t feel significantly better than Cursor’s $20 plan.
English
0
0
2
85
@levelsio
@levelsio@levelsio·
I still believe the Windows File Explorer is superior to MacOS's weird Finder
@levelsio tweet media
English
885
197
8.7K
695.1K
erel vanono
erel vanono@ErelVanono·
@FatherPhi from cost perspective Claude code isn't more expensive?
English
0
0
1
1.9K
Phi
Phi@FatherPhi·
Just last night I had it run for over 7 minutes straight solving a super tedious but easy code change, almost 40 tool calls, cursor would’ve stopped at 25
Phi tweet media
English
14
4
256
68.4K
Phi
Phi@FatherPhi·
after .5 seconds of using Claude Code
Phi tweet media
English
282
265
8.2K
1.2M
erel vanono
erel vanono@ErelVanono·
@ThePrimeagen The market really need paid AI solution for terminal experience.
English
0
0
0
290
ThePrimeagen
ThePrimeagen@ThePrimeagen·
please Cursor, keep supermaven alive and update it regularly. i'll pay you money i'll get a tattoo of it on my body i'll tell my friends about it i'll tweet about it i'll make yt videos about it i need it i need it i want it pls pls pls kthxbai
English
36
31
970
117.6K
trash
trash@trashh_dev·
“why do you rebase instead of merge?”
trash tweet media
English
115
145
4K
316.5K
erel vanono
erel vanono@ErelVanono·
I’ve been using Cursor’s background agent, and it’s been a game changer for my workflow. It’s not perfect for complex tasks, but for straightforward ones, it’s amazing. I just delegate the easy stuff while working on a feature and within minutes, I get a PR to my side branch🚀
English
0
0
2
92
erel vanono
erel vanono@ErelVanono·
@sinore69 Not sure this actually reduces verbosity. The whole point of try/catch in other languages is to group multiple method calls under one error handler. Doing that here feels more verbose, not less Maybe I am missing your intentions in that case
English
0
0
0
66
sinore
sinore@sinore18·
Go team has officially stopped considering new syntax proposals for error handling if err != nil is here to stay. What’s your take on a try/catch style syntax that doesn't change behavior, just reduces verbosity? #golang #programming
sinore tweet media
English
25
2
45
9.1K
erel vanono
erel vanono@ErelVanono·
@thdxr My concern is that, as far as I understand, they’re using SQLite to query context for the LLM. I’m not sure that’s enough to deliver an experience comparable to other available IDEs. I’d really like to see a product like this backed by a proper backend
English
1
0
0
175
dax
dax@thdxr·
this looks so good
dax tweet media
English
41
9
604
85.6K
erel vanono
erel vanono@ErelVanono·
@thdxr I have an idea let’s write book about the new field, maybe “clean AI” would be good name 😂
English
0
0
0
10
dax
dax@thdxr·
yeah we need to invent a brand new field that focuses on organizing lots of code in ways that have hierarchy and patterns that can be managed by intelligence that is limited eventually you might see people specializing in this field and companies hiring for it
Garry Tan@garrytan

AI coding agents hit a wall when codebases get massive. Even with 2M token context windows, a 10M line codebase needs 100M tokens. The real bottleneck isn't just ingesting code - it's getting models to actually pay attention to all that context effectively.

English
71
37
840
92.5K
erel vanono
erel vanono@ErelVanono·
@pnkj747 I believe Python is widely used in data science primarily because of NumPy, which provides high-performance mathematical operations and supports many essential features required by these models. Most modern ML and AI frameworks are built on top of NumPy
English
0
0
0
40
Pankaj
Pankaj@pnkj747·
If Python is so slow, why is it used in AI and data science everywhere?🤔
English
154
33
581
169.8K
typedfemale
typedfemale@typedfemale·
become terrifying to work with
typedfemale tweet media
English
157
245
6.3K
446.4K