HackerTrans
TopNewTrendsCommentsPastAskShowJobs

swe_dima

807 karmajoined 12 tahun yang lalu

comments

swe_dima
·11 jam yang lalu·discuss
the writing is on the wall for it, I have switched to gemma-4-26b-a4b.

At least in benchmarks, it scores higher and is faster.
swe_dima
·4 hari yang lalu·discuss
"You were the chosen one! You were supposed to destroy them, not join them!"
swe_dima
·7 hari yang lalu·discuss
if I travel and expect to be working I now bring a CO2 meter
swe_dima
·9 hari yang lalu·discuss
If agent is asking a question it means it stumbled on something tricky and often an important architectural decision has to be made.

Even if I am staring at the screen at the time, it's often impossible to study code and make a judgement call that quickly.

My point being, even for opt-in it can be too short - highlights how little consideration they put into it, just releasing slop in production.
swe_dima
·9 hari yang lalu·discuss
Is it possible to use their subscription pricing with Opencode?
swe_dima
·10 hari yang lalu·discuss
many years ago when Hanami was just getting popular I remember doing benchmarks against Rails when it comes to SQL and was unpleasantly surprised when Rails' ActiveRecord ended up being much faster, despite "speed" being advertised as one of the advantages :-(
swe_dima
·11 hari yang lalu·discuss
Not sure what niche it's going to occupy: too expensive for it's intelligence category.
swe_dima
·15 hari yang lalu·discuss
Pleasantly surprised that it costs as GPT 5.5, thank god for the competition.
swe_dima
·17 hari yang lalu·discuss
I found Ruby LLM to be surprisingly good - in terms of usability it's close to Vercel's AI framework.

It tries to strike a balance between working out of the box and being flexible... which has its challenges, still nice overall.

One big real-life pain I experienced is that caches don't always work, e.g. for xAI, since it only supports completions API and thought signatures are returned wrong.
swe_dima
·bulan lalu·discuss
Love Vite, but always felt sorry for them because it was not clear how they can make money, the whole VoidZero thing felt like a stretch.

It's one of those things that always stopped me from building cool tools - you have to make a living somehow.

So I am happy for the team of builders that they were able to receive the deserved payout and sustainability.
swe_dima
·bulan lalu·discuss
I still don't understand how people use git worktrees with Docker. You need a full database and etc. For me it's simpler to have multiple checkouts.
swe_dima
·2 bulan yang lalu·discuss
the video on this post looked like a Tiktok - everything is jumping, I could barely grasp what was going on
swe_dima
·2 bulan yang lalu·discuss
You may remember the argument that you can build an AI app and it continues to improve as models improve and costs go down?

Well, looking at OpenAI / Google / Anthropic we see crazy cost increases, such that it might invalidate your unit economics.

Cheering for Chinese models!
swe_dima
·2 bulan yang lalu·discuss
Flash family but costs like a Pro. $9 vs $12 for output.
swe_dima
·2 bulan yang lalu·discuss
Is it basically about giving as much data as possible to insurance companies?
swe_dima
·4 bulan yang lalu·discuss
fair enough, thanks for elaborating
swe_dima
·4 bulan yang lalu·discuss
Maybe my pockets are not deep enough, but I completely fail to understand the value proposition of iPad Air vs the regular iPad. If you want something powerful or big - go with Pro, if not - choose the "regular", much cheaper.

What am I missing?
swe_dima
·5 bulan yang lalu·discuss
Since everyone is showing their flow, here's mine:

* create a feature-name.md file in a gitignored folder

* start the file by giving the business context

* describe a high-level implementation and user flows

* describe database structure changes (I find it important not to leave it for interpretation)

* ask Claude to inspect the feature and review if for coherence, while answering its questions I ask to augment feature-name.md file with the answers

* enter Claude's plan mode and provide that feature-name.md file

* at this point it's detailed enough that rarely any corrections from me are needed
swe_dima
·5 bulan yang lalu·discuss
Definitely my experience as well.

Another dimension to this is native vs 2nd language speakers.

For those of us who had to learn English, we put a lot of effort into grammar, while native speakers whip out half-baked sentences without a second thought.
swe_dima
·5 bulan yang lalu·discuss
I admit I have not properly read a single book in a couple of years.

These days when it comes to technical stuff I much more prefer to fill in gaps by reading articles or documentation. Technical books are so long it feels like authors are paid by words.

And when it comes to fiction I have really leaned into audiobooks. My eyes are too tired from computer work, and I can combine audio with other activities like jogging or cooking.

There are some "technical" audiobooks as well, but only a small category of technical books makes sense in the audio format.