Johnny Ho

126 posts

Johnny Ho banner
Johnny Ho

Johnny Ho

@randomjohnnyh

Cofounder, CSO @perplexity_ai. Former high frequency trader, competitive programmer. Think fast, build faster.

New York, NY Katılım Aralık 2011
259 Takip Edilen139.4K Takipçiler
Johnny Ho
Johnny Ho@randomjohnnyh·
We’re open sourcing a benchmark and full evaluation harness for deep and wide research. Wide research is a particularly valuable feature for researching hundreds to thousands of entities across the web (or internally), building spreadsheets, and more. Hope this helps others build similar search and research evals for their own organizations.
Perplexity@perplexity_ai

We’re open sourcing WANDR. WANDR is an internal benchmark we built and used for building deep and wide research capabilities inside Perplexity Computer. research.perplexity.ai/articles/wandr…

English
3
1
11
1.9K
Johnny Ho
Johnny Ho@randomjohnnyh·
Memory is now organized by Concept, Entity, and Workstream by Computer's Brain. Computer Brain is a workspace agent automatically organizes your knowledge into a wiki and graph. Try it at perplexity.ai/computer/memory
Johnny Ho tweet media
English
2
0
8
2.7K
Johnny Ho
Johnny Ho@randomjohnnyh·
We've been working on a quality sprint at Perplexity, with a few nice quality of life improvements. Now available: account switching. Try it "Account > Add account" on either Mac or mobile. Web coming soon as well. Keep your work and personal environments separate but only a click away.
Computer@AskPerplexity

The Perplexity Mac app now has an account switcher. Switch between your saved accounts, add a new one, or manage them from the sidebar profile menu or the account settings menu. Live now in the Mac app for all Computer users.

English
2
1
19
3.7K
Johnny Ho
Johnny Ho@randomjohnnyh·
More efficient, less expensive models. Which will in turn lead to less stressful daily work with agents and more trust that they will accelerate and empower us rather than add additional unnecessary overhead. The future is bright for human creativity. Less compute spent on slop and more time on curiosity.
Aravind Srinivas@AravSrinivas

Imagine a fable 5 quality model that’s 3-4x less expensive in less than 6 months. And an Opus 4.8 grade model that can run on a local device in less than 12 months. Greater than 50% chance that these events will happen. Worth keeping in mind when you make predictions about the future.

English
1
1
21
6.8K
Johnny Ho
Johnny Ho@randomjohnnyh·
GLM 5.2 is an ideal model to use as an orchestrator. Computer is now cheaper and faster while maintaining capabilities on difficult tasks. Kudos to the open source community and our post training team.
Perplexity@perplexity_ai

We're releasing a research preview of a new orchestrator model in Perplexity Computer. The model is an adapted version of GLM 5.2, post-trained for the Computer harness. It delivers near-frontier performance at 0.344x of the cost of Opus.

English
2
5
46
11.1K
Johnny Ho
Johnny Ho@randomjohnnyh·
Quality of life improvement: visualize all your skills and commands in a nice command menu by typing / Deep research is now built-in to Computer too so truly everything is Computer.
Johnny Ho tweet media
English
2
1
24
3.7K
Johnny Ho
Johnny Ho@randomjohnnyh·
Brain is a new take on memory that automatically creates a knowledge graph out of the context you work on every day in Computer. Less context switching and fewer turns are needed to search for the information you need. The right context should be automatically found for you no matter where you start your work. Available as research preview today.
Perplexity@perplexity_ai

Introducing Brain in Computer. Brain is a continuously learning memory system. Every task on Computer plugs into a context graph built by Brain. It makes Computer more stateful with every run. Available as a research preview for all Perplexity Max subscribers.

English
10
4
105
12.7K
Johnny Ho
Johnny Ho@randomjohnnyh·
Congrats to @Zai_org on the accomplishment! Open weights continues to advance.
Z.ai@Zai_org

Introducing GLM-5.2: Frontier Intelligence, Open Weights - Significant improvements in coding and agentic tasks - Strong long-horizon capabilities with a 1M context window - Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong balance between performance and token efficiency - MIT-licensed open weights - Same API pricing as GLM-5.1 Tech Blog: z.ai/blog/glm-5.2 Weights: huggingface.co/zai-org/GLM-5.2 API: docs.z.ai/guides/llm/glm… Coding Plan: z.ai/subscribe Chat: chat.z.ai

English
1
3
29
3.5K
Johnny Ho
Johnny Ho@randomjohnnyh·
Grounded in actual financial metrics and news, Perplexity is the best AI at researching in-depth and realtime information about SpaceX's IPO
Johnny Ho tweet media
English
1
1
6
3.3K
Johnny Ho
Johnny Ho@randomjohnnyh·
Search as Code is state of the art on all of the top agentic search benchmarks. Perplexity's Agent API is the only API that supports Search as Code today. Learn more here: docs.perplexity.ai/docs/agent-api…
Johnny Ho tweet media
English
0
0
3
2.1K
Johnny Ho
Johnny Ho@randomjohnnyh·
Search as Code means that search can be fully integrated end to end into any agent. Search with Perplexity is composable with any CLI or file format, meaning more efficient communication, fewer agent round trips, and improved latency/cost.
Perplexity@perplexity_ai

Introducing Search as Code, our new search architecture for AI agents. It writes Python that calls our search stack directly, instead of looping through function calls one at a time. Available in the Perplexity Agent API, and now default in Computer. research.perplexity.ai/articles/rethi…

English
9
5
61
12.2K

Keşfet

@apostraphi @perplexity_ai @Zai_org @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates