ทวีตที่ปักหมุด
I made this Super Pang game in just 3 prompts with GPT-5.4. This model is really good at generating games.
English
Dominik Filkus
5.5K posts
Dominik Filkus
@DominikFilkus
⚙️ Tech lover | 🖥️ Software Engineer | 🕹️ Retro games to AI innovation | 🧠 Tesla fan
เข้าร่วม Kasım 2015
444 กำลังติดตาม634 ผู้ติดตาม
I have not managed to build a decent position so I am thinking about this strategy:
If it preruns => no action
If it pops after earnings => quit and reinvest the money elsewhere
If it drops on earnings, I could accumulate but this will depends on numbers.
What do you think about this?
English
@XFreeze I honestly do not know anybody who is using grok.
English
Grok 4.3 is sitting in the top 7 with literally just 500B parameters. The lowest size by far
Meanwhile, every other model competing at this level is between 1T to 6T parameters
It's not just small. It's also the most intelligent, fastest, and lowest-hallucination model in its class....all while being one of the cheapest to run
xAI built the most efficient frontier model on the planet
Artificial Analysis@ArtificialAnlys
xAI has launched Grok 4.3, achieving 53 on the Artificial Analysis Intelligence Index with improved agentic performance, ~40% lower input price, and ~60% lower output price than Grok 4.20 The release of Grok 4.3 places @xAI just above Muse Spark and Claude Sonnet 4.6 on the Intelligence Index, and a 4 points ahead of the latest version of Grok 4.20. Grok 4.3 improves its Artificial Analysis Intelligence Index score while reducing cost to run the benchmark suite. Key Takeaways: ➤ Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309 v2: it scores higher on the Intelligence Index while costing less to run the full benchmark suite. Grok 4.3 costs $395 to run the Artificial Analysis Intelligence Index, around 20% lower than Grok 4.20 0309 v2, despite using more output tokens. This makes it one of the lower-cost models at its intelligence level ➤ Large increase in real world agentic task performance: The largest single benchmark improvement is on GDPval-AA, where Grok 4.3 scores an ELO of 1500, up 321 points from Grok 4.20 0309 v2’s score of 1179 Grok 4.3, surpassing Gemini 3.1 Pro Preview, Muse Spark, Gpt-5.4 mini (xhigh), and Kimi K2.5. Grok 4.3 narrows the gap to the leading model on GDPval-AA, but still trails GPT-5.5 (xhigh) by 276 Elo points, with an expected win rate of ~17% against GPT-5.5 (xhigh) under the standard Elo formula ➤ Grok 4.3’s performs strongly on instruction following and agentic customer support tasks. It gains 5 points on 𝜏²-Bench Telecom to reach 98%, in line with GLM-5.1. Grok 4.3 maintains an 81% IFBench score from Grok 4.20 0309 v2 ➤ Gains 8 points on AA-Omniscience Accuracy, but at the cost of lower AA-Omniscience Non-Hallucination Rate of 8 points, so Grok 4.20 0309 v2 still leads AA-Omniscience Non-Hallucination Rate, followed by MiMo-V2.5-Pro, in line with Grok 4.3 Congratulations to @xAI and @elonmusk on the impressive release!
English
@alojohhardcore AMD is my 4th biggest position now. I have trimmed it multiple times and reallocated the cash to others but not sure how long this CPU narrative can take and what it means long term. Any recent toughs on this topic ? Could the sentiment change permanently because of this?
English
@ParthJadhav8 @Scobleizer I can’t even close the agent sidebar because all the time they are just changing hotkeys and layouts. I try to use neovim or vscode instead as much as possible. Fortunately I code less manually and living inside terminals so I rarely need an editor.
English
This is crazy !!
Cursor now has built a Kanban board where you can just drop in tasks and the agent will pick those up and complete them.
English
@alojohhardcore @jonweitzmann This HC channel is also a gift.
English
@felixmacboy @PulotRai @alojohhardcore Depends on the broker tbh. I have two brokers, one is IB, other is a local one but with that, transacting is not that easy in pre and after. IB is easy to use.
English
I think the root of the solution is an overall mentality change. I have heard arguments about how facts are more important than feelings in many areas of life and this mentality dramatically helps with making decisions. I think once you want to follow data rather than intuition or cult thinking, there will be possibilities everywhere. Of course since we are not psychos and do have emotions, it may be difficult but a clear head is such a great skill we can all develop.
English
@alojohhardcore Wild. He is a young, smart guy who saw the light after watching one of your deep, deep, deep dives into Tesla, but the slightest attempt at going against the grain is met with apathy and downright ridicule.
English
Dominik Filkus รีทวีตแล้ว
Some risks with agent wallets like this:
- Prompt injection: A clever (or malicious) input tricks the agent into requesting/approving fraudulent purchases.
- Hallucination errors: Agent misreads intent and repeatedly buys the wrong product or overpays.
- Approval fatigue: Users get spammed with micro-requests and start rubber-stamping risky ones.
- Compromised agents: If the agent's code or hosting gets hacked, it could drain funds before you notice.
Solid security helps, but AI agents aren't perfect yet.
English
Today, we’re launching the @link wallet for agents. It lets you securely empower agents to spend on your behalf. Your payment credentials are never exposed and you approve every purchase.
link.com/agents
English
@json717 @alojohhardcore Kevin is not worth the time. I watched him in the past and he has a terrible track record. He often pumped stocks like Enphase which turned out to be a financial disaster. He is a great showman but that's all.
English
@alojohhardcore AJ, I'm wondering if you saw this video. Kevin brings up some really good points.... Would love to hear your thoughts...
youtube.com/watch?v=4bSbwz…
YouTube
English
@M123dTeagan @TeslaXplored My Tesla bear/base/bull case is
$1/$500/$50,000,000.
Can I have a cookie if I am right with any of these in 2 years?
English
Good analogy about developers. In my experience, top tier developers don't only work 9–5, they work on side projects, are always open to learning, and enjoy tech discussions in their free time. They genuinely enjoy it and like the challenges.
9-5 developers are usually not the best and that mindset alone isn't enough to reach the top tier. Enthusiasm can beat talent in many cases, programming isn't necessarily rocket science but it depends on the field.
You don't need to be a geek but it helps a lot.
English
@alojohhardcore I have started wondering how to prepare for the event that Trump loses the midterms or the Democrats take the lead in the next session again. Do you have/work on a strategy for these scenarios?
English
It was on the Stripe page where apps usually redirect. The card was disabled on the UI with a message saying I couldn't use it because the payment had been declined, and I should use another payment method. So I edited the payment info, removed that card, added it again, and the payment was successful. It's like the business logic just omits this particular scenario. It was interesting because this insufficient money situation had happened before and that time I was able to use the same card without needing to re-add it.
English
@DominikFilkus Can you tell us where you were trying to make this payment? After it got declined initially, Were you asked to add the same card details again, or did you use a different card to complete the transaction? twitter.com/messages/compo…
English
Stripe works so well! (Sarcasm.) I recently failed one of my subscriptions because I had an insufficient amount of with that card.
No problem. I moved some cash there but I couldn't use that card again because Stripe says the payment was declined so I had to remove and add the card again 🤦♂️
Software engineering is dead! (No.)
English
Wow, you can now not only order a C64 Ultimate but also preorder the C64CU from the @commodoreofcl site.
Look how beautiful this Founders Edition is! 😍
English