AI tools for traders and creators, tested on real tasks. What it does, what it costs, what it gets wrong. Built by @enhale11. https://teapawt.net/t.co/OG29rinKRQ
Florida, United States
Joined January 2022
- Tweets11.7K
- Following1.7K
- Followers1.5K
- Likes10.7K
Pinned Tweet
This account is back with a new focus: AI tools for traders and creators. Each post tests one took on one real task. What it did, how long it took, what it got wrong.
Tool directory: https:// workswithai.pages.dev
Markets and AI for business: Enhale11
GitHub now lets you manage Copilot code review settings through the REST and GraphQL APIs, from enterprise policy down to a single repo.
Pro through Enterprise. Review overviews are more compact, and agent sessions show AI credits used per review.
github.blog/changelog/2026-1…
Google Jules stops new sessions on Nov 11, then goes read-only until it winds down Dec 31.
Export settings, memories, and session history with the Jules CLI before then. Your GitHub branches and PRs stay put.
jules.google/docs/changelog/…
Qoder's Qoduck stealth model is free daily for individual users from Nov 7.
Within the allowance it bills at 0.0x Credits (was 0.5x), even at max thinking. Teams/Enterprise and the IDE, JetBrains plugin, and CLI aren't covered.
docs.qoder.com/events/qoduck…
DuckDB's Claude Code plugin routes data jobs through the DuckDB CLI instead of guessed Python scripts.
Nine skills cover read, query, convert, S3 explore, and spatial. Install: /plugin install duckdb-skills@claude-plugins-official
duckdb.org/2026/09/16/duckdb…
Anthropic's engineering post on agent memory: "dreaming" checks notes against session transcripts and rewrites mistakes.
Claude Code only injects the first 200 lines or 25 KB of MEMORY.md each session. Keep that header as an index, not a dump.
Claude Code 2.1.317 asks before a web or shell call sends your account email to the network.
That landed in the Nov 3 release notes. Update if you're on an older build, and leave the privacy check on.
github.com/anthropics/claude…
Oumi's FAiE turns a plain-language goal into a model or agent plan, then waits for your OK before jobs that spend money.
Pro is $25 a month on oumi.ai/pricing, then pay-as-you-go for training and evals. Leave Auto-approve off until you know which actions it starts.
Gemini Omni 1.2 Flash is GA on Google's Gemini API.
Native clips go to 15 seconds (was 10). Price is $0.07/sec for 720p, 30% below 1.1. You can edit the same clip in chat instead of regenerating.
Google still notes weak spots: fast sports, rapid script changes, complex physics.
Anthropic's Ultra plan is $200 a month on claude.com/pricing.
What it adds over Max ($100): no 5-hour limit for verified accounts, 10x Pro usage each week, and $200 in monthly API credits instead of $100.
Worth it only if Max's 5-hour wall keeps stopping long agent runs.
webAI open-sourced TwIL-8B, an 8.8B-parameter model small enough to run on a laptop. Weights are on Hugging Face.
Its own numbers: 93% of frontier performance on coding, 92% on instruction following.
Those are the company's tests. Try it on notes you'd keep offline.
Cognition released SWE-2 Flash, a smaller coding model for its Devin agent.
Its claim: 5x cheaper and 3x faster than SWE-2, at about 200 tokens a second. It's free in Devin Desktop and CLI through November on Pro, Max and Teams.
A good month to try it on code review.
Tencent's Hy4 is out. The Hunyuan model now reads images, charts and screenshots, and can use a computer.
Listed price: $0.834 per million input tokens, with a 1-million-token context. It's on Tencent Cloud and OpenRouter.
First test worth running: a chart you already know.
Revolut reportedly moved its 2,000 engineers onto a self-hosted open-weight model, GLM, at about 30 cents per million tokens, about a tenth of its old bills.
The part worth copying: tough tasks still go to frontier models.
Cheap model for daily work, paid one for hard cases.
Agent37, a YC-backed startup, hosts AI agents in sandboxes billed per minute and paused when idle.
Its founder's comparison: $4.76 a month for a 2 vCPU, 4 GB setup vs $249 on Vercel.
Vercel's CEO disputed it, since the two bill CPU differently. Price your own idle time first.
If you make AI video, StreamEval is worth a look.
It's an open benchmark that runs 38 video models through live sessions of 15 to 300 seconds while the instructions change.
Its finding: sharp image quality doesn't mean the model holds up mid-session.
streameval.ai
GPT-6.1 Terra, OpenAI's model for computer use and agent tasks: $1 per million input tokens, $5 per million output, half the price of GPT-6.1 Sol.
Early benchmarks: it trails at lower effort and only matches top models at max effort.
Price it at the effort you'll run.
Spotify is reportedly building a paid AI remix and cover tool, aimed at early 2027.
Reported price: $4 to $5 a month on top of a normal plan.
The limit creators should note: it only works on songs from artists who opt in, and royalties go back to those artists.
Got the Claude Code extension in Cursor around Oct 29? Check the version.
2.1.312 and 2.1.313 were tampered with and up for about 11 hours, reaching up to 5,000 users.
Updating may not clean it up. Check for unknown SSH keys and startup entries, and rotate credentials.
Webflow's Claude connector lists 24 tools, covering pages, CMS, styles, forms and the sitemap.
What Webflow's docs say it can't do: start a new site, or create new localized CMS items. Image uploads need the Designer open with the Bridge App.
Plan limits aren't documented yet.