inmydata

101 posts

inmydata banner
inmydata

inmydata

@inmydata

80% of what your organisation knows lives in people's heads. We capture that expertise, connect it with operational data, and build AI that improves with use.

Edinburgh, UK Katılım Ağustos 2017
22 Takip Edilen9 Takipçiler
inmydata
inmydata@inmydata·
In seven days, GitHub repriced annual Copilot and Anthropic A/B tested Claude Code out of the $20 plan. Flat-rate billing has broken under #agentic load. The fix is in the architecture, not the invoice. inmydata.ai/blog/github-ch…
inmydata tweet media
English
0
3
3
75
inmydata
inmydata@inmydata·
Great to see Scotland's tech community gathering again at #ScotSoft2026. Our CTO Nick Finch will be on stage exploring how agentic AI is reshaping software UI, and what that means for the next wave of products built here. See you at the EICC in September.
ScotlandIS@scotlandis

Meet our next speaker for #ScotSoft2026 - Nick Finch! Nick is CTO and co-founder of Edinburgh-based @inmydata, and will explore how #agentic #AI is transforming traditional #UI and the opportunity this poses. Earlybird tickets on sale now! scotlandis.com/scotsoft-2026/

English
0
3
4
80
inmydata
inmydata@inmydata·
#Opus 4.7 deprecated temperature, top_p, top_k, and extended thinking budgets In their place, one new parameter. task_budget Stop tuning. Start bounding how much thinking the model can afford For some, a config change, for others a rewrite. inmydata.ai/blog/opus-47-t…
inmydata tweet media
English
0
3
3
41
inmydata
inmydata@inmydata·
Organisations struggling with #AI costs are building #agents like #chatbots. Fire a prompt, get a response, worry about the bill later. That doesn't work for #agentic systems. Constraint isn't the opposite of capability. It's what makes capability viable inmydata.ai/blog/why-const…
inmydata tweet media
English
0
3
3
45
inmydata
inmydata@inmydata·
Computing solved the structured data problem fifty years ago. The other eighty percent, the expertise and judgment that actually drives decisions, has never been computerised. Until now. inmydata.ai/blog/the-knowl… #AI #RAG
inmydata tweet media
English
0
3
3
31
inmydata
inmydata@inmydata·
Chroma Research tested 18 frontier models. Every single one gets worse as you fill the context window. The bottleneck wasn't capacity. It was curation. Read what #ContextEngineering actually looks like when you're building it, not just talking about it inmydata.ai/blog/stop-tuni…
inmydata tweet media
English
0
3
3
26
inmydata
inmydata@inmydata·
Three #MCP servers in production. One serves operational #data. One serves expert knowledge via #agentic #RAG. The third lets agents request new knowledge by interviewing experts. Three types of intelligence, one protocol. Composability is the unlock. inmydata.ai/blog/mcp-won-n…
inmydata tweet media
English
2
3
3
29
inmydata
inmydata@inmydata·
We built two AI agent systems in 4 months. Work that would have taken 2 years. The secret isn't the model. It's that AI built the systems too. Data pipelines, business logic, retrieval architecture. Be AI-first in how you build, not just what you build. inmydata.ai/blog/ai-first-…
inmydata tweet media
English
0
3
3
15
inmydata
inmydata@inmydata·
The incumbents still here in 2030 won't be the ones that bolted #AI on and called it an #agent They'll be the ones who accepted the seat model is broken and competed on what you can't recreate over a weekend 30 years of the right data. That's the moat inmydata.ai/blog/the-incum…
inmydata tweet media
English
0
3
3
25
inmydata
inmydata@inmydata·
1:30pm - started building a pen testing suite 3:30pm - finished 4:00pm - first scans running 5:00pm - issues fixed, gone to the pub A commercial pen test costs £5-25k and takes weeks. Read why this pattern explains the turbulence in software stocks inmydata.ai/blog/when-ai-e…
inmydata tweet media
English
0
3
3
30
inmydata
inmydata@inmydata·
On February 3rd, $285 billion was wiped from #software stocks in a single day. By mid-February, a trillion. It wasn't a #bubble bursting. It was #AI getting good enough to do the work those companies charge for. That money didn't disappear. It moved inmydata.ai/blog/when-ai-e…
inmydata tweet media
English
0
3
3
23
inmydata
inmydata@inmydata·
#AI #slop isn’t a model problem. It’s a process failure. Treat #LLMs like vending machines and you’ll get average output. No plan. No review. No judgement. AI can generate prose. Only process creates meaning. Slop is what happens when you skip the work. inmydata.ai/blog/we-are-no…
inmydata tweet media
English
0
3
3
17
inmydata
inmydata@inmydata·
We finally got the AI assistant people wanted. Not because it answers better, but because it gets things done. That’s also why it’s dangerous. Alexa is safe because it can’t do much. @openclaw is exciting because it can do almost anything. inmydata.ai/blog/openclaw-… #AgenticAI #AI
inmydata tweet media
English
0
3
3
21
inmydata
inmydata@inmydata·
Software spent 40 years piling up menus, forms and workflows so humans could translate intent into clicks. Agents invert that. You state the outcome. They do the clicks. The UI becomes a readout not the control panel. inmydata.ai/blog/why-your-…
inmydata tweet media
English
0
3
3
13