hexmint ✌️ @hexmint

Sri Lanka Joined January 2011

Tweets

18K
Followers

902
Following

2K
Likes

43K

hexmint ✌️ @hexmint

2 hours ago

Regardless of whether Jay Shah said it, we will need an election, that part is obvious.

What is happening to our cricket? In April Daily Mirror exclusively reported saying SLC Chief Shammi Silva had been bowled out by President AKD & an interim committee will take over SLC headed by Eran Wick. - Sports Ministry denied the story - SLC denied the story - Cabinet

6 6 52 8K 3

0 0 0 9 0

View Details

hexmint ✌️ @hexmint

2 hours ago

Whenever you do a rewrite of something, try to use the strangler-fig pattern, because it's hard to catch running targets. If that's not possible, make sure you don't forget the hard learned lessons from the old codebase. Some complexity is inherent and you should only get rid of the additional implementation complexity. And if you use AI, understand that you have to use your brain too, at least slightly. You can't "hey, do this" and be okay with the output; if you do so, I guarantee you will have to throw away the whole thing sooner or later.

0 0 0 35 0

View Details

atomic.chat @atomic_chat_hq

10 hours ago

Sakana Fugu surprisingly performed near GLM 5.2 level but 17× more expensive! We gave the same prompt to 4 models: build a complete live Trader Desk with both frontend and backend components, real-time market data fetched from external APIs for 8 symbols, and a custom dark-theme UI. Outputs: Fugu Ultra — 22,225 t, $0.51 Opus 4.8 — 15,802 t, $0.31 GPT-5.5 — 11,474 t, $0.26 GLM 5.2 — 13,677 t, $0.03 Fugu created the most polished and feature-rich trading desk in the run. GLM 5.2 was very close behind, with a similarly complete multi-panel interface and live data, but at a much lower cost. Opus and GPT also performed well, delivering solid results with a better balance between quality and cost

Sakana AI @SakanaAILabs

a day ago

Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API. Our ‘Fugu Ultra’ model matches the performance of Fable and Mythos, delivering frontier capability without the risk of export controls. Try it: sakana.ai/fugu 🐡

1K 6K 36K 23.1M 29K

58 139 1K 286K 740

View Details

hexmint ✌️ @hexmint

17 hours ago

Just realized "trader" and "retard" are anagrams

0 0 1 18 0

View Details

hexmint ✌️ @hexmint

23 hours ago

Cats get asthma? I thought I got asthma because I was with cats all the time

Cats with Aura 😺 @catwithaura

a day ago

Giving a cat an inhaler for asthma

35 178 4K 129K 322

0 0 0 19 0

View Details

猫しか勝たん @A1P4CA

2 days ago

やけにハンドソープの減りが早いと思ったら…

101 613 5K 604K 499

View Details

hexmint ✌️ @hexmint

2 days ago

Well you're wrong. Regardless of whether someone uses AI to write their code, they should be able to take responsibility for that code, and should be able to explain the architecture and the general flow of what's happening. Some can't and they are the said lazy developers.

Xiao Ma @infoxiao

3 days ago

I just open sourced my “lazy or craftsman?” simple test.

10 41 748 51K 95

0 0 1 38 0

View Details

hexmint ✌️ @hexmint

2 days ago

@mattpocockuk What do you mean exit the smart zone?

1 0 2 421 0

View Details

BURKOV @burkov

3 days ago

For the last three days, I've been using GLM 5.2 with OpenCode instead of Codex and I don't see any difference. There wasn't any bug that GLM would fail to fix or a feature it would fail to add as requested. The only downside is that this model cannot see, so if it's simpler to explain an issue by pasting a screenshot, I would still use Codex. Otherwise, GLM would be my choice. Will continue to use it for two more weeks and, if it keeps just working, I will cancel my $100/month subscription with OpenAI. I already cancelled my Anthropic subscription and have no regrets. No moat isn't hypothetical anymore.

164 125 2K 187K 566

View Details

hexmint ✌️ @hexmint

4 days ago

Nobody serious cares even now (doesn't mean it will not be great ever)

Marc Lou @marclou

5 days ago

.@mntruell launched Cursor 8 times, and nobody cared. Don't give up.

269 605 8K 437K 2K

0 0 0 34 0

View Details

hexmint ✌️ @hexmint

4 days ago

@arunoda Awesome 😎

0 0 0 25 0

View Details

Zack D. Films @zackdfilms1

6 days ago

Why They Painted One Turbine Blade Black 🤔

61 144 5K 951K 498

View Details

NewsWire 🇱🇰 @NewsWireLK

5 days ago

🚨 Pizza Hut is set to be sold in a $2.7 billion deal after years of struggling sales and increased competition. Parent company Yum! Brands has agreed to sell Pizza Hut's global operations (excluding China) for $1.5 billion, while the China business will be sold separately for $1.2 billion. The deal comes as the iconic pizza chain faces declining sales, changing consumer habits and growing competition in the fast-food market. The transactions are expected to be completed later this year, subject to regulatory approvals.

10 26 179 22K 9

View Details

hexmint ✌️ @hexmint

6 days ago

I joined Cursor early 2024 and used it for 7 months, most of it for free. Even then it had this "compose" (or composer but not the model by the same name) mode where it allowed us to select a few files and tell what to do. I think back then it rewrote the whole files instead of replacing lines.

0 0 0 54 0

View Details

hexmint ✌️ @hexmint

6 days ago

@_merovingienne_ Same. Didn't use for months but tried.

0 0 1 27 0

View Details

hexmint ✌️ @hexmint

6 days ago

@Imesha94 This was GLM 5.1. Okay but, Claude is at least trained on visual data.

hexmint ✌️ @hexmint

2 months ago

GLM 5.1 isn't even a visual model, and it created a certain sea shell shape, in SVG, almost perfectly by reading a bloody description from the internet.

0 0 1 108 0

1 0 0 30 0

View Details

hexmint ✌️ @hexmint

6 days ago

Even without vision, it makes good SVGs, which is crazy.

Sawyer Hood @sawyerhood

6 days ago

one thing that has always puzzled me is that the glm models don’t have vision so it greatly reduces the usefulness of this

30 3 189 45K 17

1 0 0 48 0

View Details

hexmint ✌️ @hexmint

6 days ago

GLM version is more frightening too

AI/ML API @aimlapi

6 days ago

GLM-5.2 is now available on AI/ML API! We asked it and Opus 4.8 to one-shot a playable Backrooms, same prompt. Opus got the flashlight working but you can't run and can't pause. GLM-5.2 built the full mechanics. Opus: 2:14sec & $1.94 GLM: 1:08sec & $0.37 Docs, guides & setup